The Guide, which starts here in Post 1, and which continues in Post 2, is intended to be a general guide to Home Theater, HT calibration, and audio quality. Due to its roughly 250 page length, I have had to divide it into two posts. Sections I through III follow the Introduction in this post. Sections IV through VIII are in Post 2. The discussion thread, which starts after the Guide in Post 3, is a general audio/HT discussion thread.
Parts of the Guide may seem very thorough and detailed, but readers can choose the parts they wish to read at any one time. Over a period of time, however, most of the information contained in the Guide is likely to prove valuable for those who want to understand how things work, and for those who want to get the most from their HT/audio systems.
The Guide began on the Audyssey thread as a way to explain Audyssey's calibration process, and to provide guidance with respect to adding bass boosts after calibrations. But, over time, the Guide has expanded to encompass discussions of a wide-ranging series of topics involving our HT and audio systems, many of which have nothing to do with any system of room correction. The Table of Contents, and the Introduction to the Guide, provide an overview of the topics which are discussed.
* Audyssey remains the most commonly used system of room correction. Where Audyssey-related sections are involved, the basic principles of room/system interaction, of system calibration, and of room EQ, which are contained in the Guide, are believed to be generally applicable to other methods of HT calibration, and to other systems of automated room EQ.
With respect to the thread which follows the actual Guide, ideally this will be a thread where people will be comfortable discussing any HT calibration, subwoofer, or audio-related issues.
Cliff Notes:
(Tips For Getting Started With Your HT System)
Several people have suggested having some abbreviated HT calibration tips, so I have added a few simple tips prior to the actual Guide. Anyone wanting more detail on anything from room and speaker setup, to setting crossovers, to the differences between sealed and ported subs, to selecting and positioning subwoofers in a room, can find the pertinent information in the Guide itself. There are some similar abbreviated tips, to selecting subwoofers, at the beginning of Section VIII.
1. Try to position speakers and subwoofers strategically, doing a "subwoofer crawl" if necessary. Section I-B gives some good general advice on locating the speakers on the front soundstage. (The front three speakers will have the most impact on the overall sound quality for most of the frequency range.) You can Google how to do a sub crawl, or you can refer to Section VIII-E for instructions on how to perform the procedure.
2. Set the subwoofer phase at 0, and the low-pass filter (the LPF is sometimes labelled crossover) on the subwoofer at the maximum setting. Typically, the gain control on the sub (which is often labelled volume) should be set at or slightly below about the mid-way point. This varies! With some subs, the gain may need to be at 10:00 or 11:00 on an analogue dial. On others, such as some Monoprice subs, the gain may need to be at 3:00 or higher. During the level-matching process, which occurs at the first mic position, you will actually want the subwoofer(s) to be slightly above the volume that Audyssey is telling you to achieve, for reasons which are explained in #12. of the Cliff Notes. (On newer Denon/Marantz AVR's, that means that you will want your subwoofer volume to be slightly in the red zone.)
3. Allow your AVR to calibrate your audio system for you, using its automated routine. The first thing that will do is to calibrate your HT system to Dolby/THX Reference (which will correspond to 0.0 MV) , and it will provide a common basis for comparing listening levels and subwoofer boosts. That calibration process will insure that all of the channels are playing the same volume levels at the MLP (main listening position) and that all of the sounds will arrive at the same time. Those equal volume levels are also essential in order for the room correction process to occur. Room EQ only works to minimize peaks and dips in the sound if all channels are playing at the same volume level.
4. The way that Audyssey and other forms of auto-calibration work is that speaker levels and distances are set from the first microphone position, which is always defined by your AVR as the MLP. As noted above, that ensures that all channels play the same volume and that all sounds arrive at the same time. Subwoofers have internal processing, through their own internal amplifiers, which delays the arrival of the sound, and wireless subs have even more delay. So, their distance settings will not correspond exactly to their physical distance from the MLP. The distances will be greater than the physical separation from the MLP. This is normal, and the distances should be left as they are, unless there is some other specific reason to change them.
5. The second thing that the calibration will do is to set EQ filters, for all of the channels, to reduce some peaks and dips in the frequency response caused by the interaction of your transducers (speakers and subwoofers) with the room. Those random peaks and dips in volume, at different frequencies, interfere with the quality of the sound we hear. With subwoofers, boomy, one-note bass is often the result of a random peak. After room EQ, the bass may sound smoother, but correspondingly less impactful, until the subwoofer volume is increased. Increasing the subwoofer volume after calibrating is normal, as explained in Note 10.
6. It is important to understand that two different things are occurring during a calibration. The initial calibration process ensures that equal volume levels, from all of the channels, will arrive at the MLP at the same time. And it calibrates the audio system to a "Reference" standard. The room EQ process (that is also part of the calibration) sets filters, for all of the channels on an individual basis, in an effort to improve the overall sound quality in the room. To do that, it needs to start with equal volume levels. From those equal volume levels for each channel, room EQ will add or subtract volume at specific frequencies, to get as close as possible to the target volume of 75dB. Where room EQ is successful, a measured frequency response will show a somewhat flat line from the lower frequencies to the higher frequencies. (Once random peaks and dips are removed, listeners still have the option to tilt that somewhat flatter line toward their particular listening preferences.)
7. It is also important to emphasize that all channels are EQed individually, in relation to the room, and not in relation to each other. The speakers are not EQed with respect to each other or to the subwoofers. Each individual channel is EQed with respect to the room and the MLP. All subwoofers in a system are treated as a single channel, even if there are two Sub out's in the AVR. The two sub outs may allow the AVR to set separate volume levels and distances, based on the specific subwoofer positions in the room. Whether or not a particular AVR can set separate volume levels and distances will depend on the model. But, irrespective of whether there are separate sub outs, all subwoofers are EQed together, as if they were a single sub.
8. There is a model calibration procedure shown in Section I-C which may help you to achieve an optimum calibration. It has a diagram of potential microphone positions which seem to work well for many listeners. Starting with a good speaker setup, which is addressed in Section I-B, and a good calibration, can take some effort. But, doing those things can make an audible difference in the resulting sound.
9. As noted, when an AVR calibrates your audio system, all of your channels including your subwoofer(s) will typically be set to play the same volume at the main listening position (MLP), and you will be listening to Reference volumes when you are at a listening level of 0.0 MV (master volume). Most people probably listen at an average volume of about -15 MV to -20 MV. Individual listening volumes, however, can vary much more widely than that. (Your master volume is the only AVR setting that will be unchanged after an Audyssey calibration. It will still be wherever you had set it prior to running Audyssey.)
10. Since all of the channels are now playing equal volumes, and since we don't hear low-frequencies as well as other frequencies, after calibration most people will need to add more bass to their audio systems. That is particularly the case when we listen at below Reference volumes, where bass frequencies were designed to be in better equilibrium with frequencies in our normal hearing range of about 500Hz to 5,000Hz. As we drop below Reference (0.0 MV), bass frequencies drop-out of our hearing much more quickly than other frequencies do. It is fairly typical to add +3dB to +6dB of subwoofer boost on top of Audyssey's DEQ, and even more than that if DEQ is disabled.
11. We can compensate for the audible reduction in bass by turning-up the volume of the subwoofers, and how much volume to add is strictly a user preference issue. To state this in a different way, after an Audyssey calibration, very large subwoofers will be playing at exactly the same volume level as very small subwoofers would be, since all of the channels in a calibrated HT system are level-matched to play the same SPL at the main listening position. In order to use the greater available output of more powerful subwoofers, it is simply necessary to turn-up their volume after the calibration.
12. After running Audyssey, it is generally desirable to add most of your subwoofer volume increase with your subwoofer gain control, while not letting your AVR sub trim go above about -5, in order to avoid clipping the pre-out signal coming from your AVR. That is especially important at master volume levels above about -15. Typically, it is a good idea to raise the gain on the subwoofer high enough to achieve a trim level of about -9 to -11.5 during the initial level-matching process. (Gain and trim are inversely proportional during the level-matching process. Raise the sub's gain and the AVR trim goes down, and vice-versa.)
The lowest trim level that you should use with Denon/Marantz is -11.5, because the trim controls only go down to -12. If you are at a trim level of -12, you won't know whether your trim level actually should have been even lower than that. In some cases, it might have needed to be much lower than -12 if the trim levels could go that low.
An expedient way to set trim levels during a calibration is to just run three mic positions and tell Audyssey to calibrate. Then, you can look at the trim levels that Audyssey set. Once you get the AVR trim results you want, you can do a full 8-point calibration with the higher versions of Audyssey, or with 6 points for the lower versions. To make it even quicker, you can calibrate only for your front speakers and your sub(s). Just go into your Speaker Configuration menu to add or subtract channels when you use this approach. And, you can keep the microphone in the same spot for those three sets of sweeps. You are only setting trim levels to coincide at the MLP anyway, during the level-matching process.
It will probably take a volume level of about 78dB to 80dB, instead of Audyssey's default 75dB, to achieve low trim levels in about the -9 to -11.5 range. Where XT-32, with SubEQ is used, that will put the subwoofer volume in the 'red zone'. That's perfectly fine, just tell Audyssey to continue. After running Audyssey, we can conveniently raise the AVR trim to about -5 or -6, with our AVR remotes, and we can continue to increase the subwoofer gain if we want even more bass than that. Section II explains the best ways to use the subwoofer gain in some detail, and explains in greater detail why it is generally advisable to keep AVR subwoofer trim levels well in negative numbers.
* If a listener doesn't want to run Audyssey again, but has a subwoofer trim level that is higher than the Guide is recommending, the following procedure is perfectly acceptable. There is nothing wrong with simply lowering the AVR trim level, after a calibration, and then raising the subwoofer gain level to compensate for that. So for example, rather than doing a new calibration, a listener could lower the trim level to about -5 or -6 and then raise the subwoofer gain until the bass seemed loud enough. Ultimately, as with the master volume, everyone will just add or subtract bass in accordance with his own personal preferences.
13. It is important to understand that the amplifier connected to our subwoofer output in the AVR is not like the amplifier in our AVR which powers our speakers. The amplifier connected to our sub out is only intended to power on the subwoofer from Auto (Sleep) mode, and to make convenient incremental changes to our subwoofer volume. Our subwoofers have internal amplifiers which are needed for really serious increases in subwoofer volume. And, we can increase our volume with our gain controls as much as we want to, after a calibration, and use our AVR remote to make smaller up or down changes while keeping our AVR trim well in negative numbers.
There is a short article which explains this in more detail here:
14. Some Denon AVR's have a feature called Subwoofer Level Adjust. If a Denon AVR has an On/Off control for that feature, it should typically be turned off, and any volume adjustments should be made either with the subwoofer's gain control, with the subwoofer trim control in the Audio menu, or with the trim control in the Speaker: Manual: Test Tone area of the AVR. Again, in making any subwoofer trim adjustments, it is desirable to keep the trim levels at about -5 or lower. To add more subwoofer boost in excess of a -5 trim setting, it is always perfectly acceptable to use the gain control on the subwoofer. (You can keep track of your initial gain setting on an analogue dial by marking that hatch mark with a small piece of tape.)
15. It is important to understand that Audyssey will measure your speakers, at their specific positions in your room, and your AVR will set preliminary crossovers in accordance with its own programming. For instance, if a particular speaker pair, or center channel is capable of playing below 40Hz, at its specific position inside a room, and at the 75dB test tone, your AVR will set the speaker to "Large". Or a 40Hz or 60Hz crossover may be set, if speakers can't go quite low enough for an initial setting of Large. Those aren't actually recommendations! They are just observations, based on the measured response of your speakers. After an Audyssey calibration, it is advisable to reset crossovers as suggested below.
16. If there is a subwoofer in the system, speakers should typically be set to Small, and crossovers should typically be set at 80Hz or higher. (The 75dB test tone that our AVR's use isn't actually very loud. As we go up in volume, the speakers' low-frequency capabilities will degrade. Raising crossovers transfers more of the low-frequency demand to your more powerful subwoofer.) The LPF of LFE, in the AVR, should typically be set to 120Hz (which is usually the default setting in our AVR's). There are exceptions to these settings, which are explained in Section III, but this is typical best practice advice for starting-out with an HT system.
(As of October 2023, I have been informed that new Denon/Marantz AVR's no longer have a Large/Small setting. Listeners simply set crossovers for the various speakers.)
17. It is always acceptable (and often desirable) to raise crossovers from their initial calibration setting, but it is not generally desirable to lower them from wherever your AVR set them. Among other things, Audyssey will not be EQing speakers below the crossovers set during the calibration process. And, as noted above, 75dB is not very loud. As your speakers try to play peak volume levels greater than 75dB, they will roll-off faster and that will put more demand on them. It is usually better to let the subwoofers do the heavy lifting. Crossovers are explained in detail in Section III.
18. If you have Audyssey, you can add subwoofer boosts on top of Dynamic EQ, or turn off DEQ and add your own subwoofer boosts. Section V explains DEQ in detail. Experimenting with it on, and with it off, may be helpful. Turning it off will probably require you to add more subwoofer volume, but for some people, it may change the sound in a positive way. There are also RLO (Reference Level Offset) settings, which are associated with DEQ, and which may be helpful to moderate its effects.
19. You can turn Audyssey off, to hear how things sound without room EQ, and then turn it back on, without changing the room correction filters that it set for any of the channels. And, you can experiment with Audyssey Flat. Audyssey (Reference) and DEQ are always the default settings after a calibration. Audyssey Flat and/or DEQ off are user preference options, and both are explained in Section V.
20. After an Audyssey calibration, you can change any settings in your AVR without affecting the room correction filters that Audyssey sets. Changing AVR settings prior to running an Audyssey calibration will not be helpful. Audyssey is designed to ignore and override prior settings when it calibrates an audio system. The only setting that will remain unchanged after an Audyssey calibration is the master volume. But now, your AVR and your audio system will be calibrated to correspond to Dolby/THX Reference, when the MV is set to 0.0.
21. If you make a significant change to a room, such as moving a speaker or a subwoofer; changing to different speakers or subwoofers; adding new subwoofers or other channels; or adding room treatments or making significant furniture changes; you should recalibrate.
22. It may be important to recognize that virtually all settings, including your master volume level, are listener-preference settings. There is no universally correct way to listen to music, or to watch movies or TV shows. Even the use of room correction, in whatever form, is a user-preference feature. Some people prefer listening without room correction, or limiting its effect to just the lower frequencies. And, that same idea applies to all of the settings associated with room EQ, or with our AVR's in general. There are default settings that may help us to get started, and there are some best practice principles which we may want to follow. But, we will all define audio quality in slightly different ways, and we will all have slightly (or profoundly) different listening preferences. Informed experimentation can be the key to discover what we really like.
23. When you try different calibrations, or different settings, remember that you will ultimately have to trust your own judgment with respect to sound quality. The Guide, and other sources, can help to explain some audio theory, and how certain features work. Those same sources can make suggestions regarding general best practice principles, and can offer options for things that listeners can experiment with. But, in the end, everyone will have to decide for himself what he actually likes.
Each of us decides for himself how much, or how little, he wants to experiment with his audio system. Some of us may just be looking for a plug-and-play approach to our HT's. That is perfectly fine too! After all, the goal here is simply to please ourselves with respect to our entertainment hobby. And, for those of us who do want to experiment, each of us also decides when he is satisfied and wants to stop experimenting. It is not unusual to stop and just enjoy our audio systems for a while, and then to experiment again weeks, months, or even years later.
The way in which you experiment and listen can be important! Take your time! Try a particular setting for several days, unless you are absolutely sure that you don't like it, before trying a completely different one. Ideally, you want to let your hearing adjust to one sound quality before trying something different. What you don't want to do is to introduce several new variables all at once, because you won't be able to separate them, and you may not have really learned anything about your own listening preferences.
You also don't want to overload your own hearing, and your brain's response to what you are hearing, by trying to pack too many changes and too many concentrated listening sessions into too short a period of time. I referred to them as "concentrated" listening sessions, but that is really the wrong word. They really need to be as relaxed and natural listening sessions as you can make them.
Your brain will tell you what you do and don't like, if you relax and give it some time. You don't have to 'concentrate' to decide whether something tastes good or not, or how much seasoning you prefer, or whether you like a particular color. Concentrating on trying to hear specific things in your sound can actually be counterproductive. Just relax and enjoy the process of experimenting, and of making gradual, incremental improvements in your sound quality.
If you listen for two or three days, your hearing will adjust somewhat to that particular sound. Then, when you do try a different setting you will have an audio benchmark with which to compare any changes in the sound. You won't necessarily have to concentrate there either. Just let the listening sessions happen naturally. If there is no audible change in the sound, then you may not need to worry about that particular setting. If you feel that a new setting has a positive effect on the sound, make a note of that setting, and of your reaction to it. Keeping track of which settings work best for you, and why they seem to improve things, will help as you continue to experiment.
If a setting seems to have a negative effect, you might cross that one off your list right away, or perhaps return to it later. Above all, be patient and take your time. Impatience will only take you in circles! There are just too many different ways to achieve improved sound quality, and too many variables, for us to try to go too fast. But, if we are patient and systematic in our experimentation, almost everyone will get to a final result that is most appropriate for that particular listener. And, we can stop experimenting, and just enjoy our audio, anytime we choose.
____
Specific sections of the Guide deal with how to setup speakers and how to perform good automated calibrations. Other sections deal with setting crossovers, and selecting and positioning subwoofers. Most of the terms used in the Guide are defined in the first two sections. Readers can choose what sections to read for a particular purpose, although some of the information does build sequentially as you go along.
* The Guide is organized into the following major sections. Each section (and subsection) is hyperlinked so that readers can go directly to that part of the Guide by clicking on it in the Table of Contents. Individual sections or subsections can also be copied and pasted, by right clicking on them in the Table of Contents, for inclusion in a post. That will enable others to go directly to the subsection to which you are referring.
Note: With the transition to the new XenForo platform, the previous hyperlinks were lost. I have added new hyperlinks to allow internal navigation within the Guide, but they currently take us to a position about three lines down from where they should. The Forum Administrator has asked the Development Team to fix that glitch, but it appears that may never happen. In any event, it it still much quicker to use the blue hyperlinks, and then to scroll back up a couple of lines.
Table of Contents:
Introduction to the Guide:
Section I: Room/System Setup and Sound Quality:
Section I-A: The Frequency Range:
Section I-B: Distortion, Speaker Placement, and Room Treatments
Section I-C: Room EQ and Calibration Techniques
Section II: Audio System Calibration and Subwoofer Levels:
Section II-A: Audyssey Calibration And Dolby Reference
Section II-B: Why We Add Bass After Calibrations
Section II-C: Where And How To Add Bass
Section II-D: Master Volume Levels And Sub Boosts
Section II-E: Gain Settings And Maximum Sub Output
Section III: Setting Crossovers:
Section III-A: Crossovers From Speakers to Subwoofers:
Section III-B: Low Frequency Effects Channel:
Section III-C: Cascading Crossovers:
Section III-D: Bass Localization:
Section III-E: LFE+Main:
Section IV: Integrating Multiple Subwoofers:
Section IV-A: Setup and Calibration:
Section IV-B: Room EQ:
Section V: Audyssey Dynamic EQ and Dynamic Volume:
Section V-A: Dynamic EQ:
Section V-B: Tone controls and House Curves:
Section V-C: Dynamic Volume:
Section VI: Audyssey Thread History of Recommended Subwoofer Trim Settings:
Section VII: Bass Frequencies, Room Gain, and The Equal Loudness Contours:
Section VII-A: Bass Frequencies and Tactile Response:
Section VII-B: Room Gain:
Section VII-C: The Equal Loudness Contours:
Section VIII: Bass Preferences, and Subwoofer Selection and Placement:
Section VIII-A: Sealed Versus Ported Subwoofers:
Section VIII-B: Comparing Subwoofer Performance:
Section VIII-C: Selecting Single Versus Multiple Subwoofers:
Section VIII-D: Internet Direct Subwoofers:
Section VIII-E: Subwoofer Placement in a Room:
Introduction to the Guide:
The most commonly asked question on many AVR and room correction threads, and on a number of subwoofer owners' threads, involves subwoofer settings. People who have new audio or home theater (HT) systems, or who have upgraded and/or added subwoofers, are naturally anxious to be able to get the most from them. In addition, there is a fairly universal perception that bass volumes sound somewhat softer after running Audyssey, or YPAO, or other systems of automated calibration. And, people are frequently curious about whether that perception is normal, and if so, about the best way to increase their bass.
The Guide was originally written to explain why it may be perfectly normal to perceive bass levels as lower, after running Audyssey or other forms of automated calibration. And, it was written to explain the best ways to use a combination of subwoofer gain and AVR trim to make bass boosts. In attempting to address issues involving subwoofer boosts, however, I have found that it is also helpful to understand some of the basic principles of HT system calibration, and their relationship to Dolby Reference.
And that, in turn, has led to discussions of how we hear bass frequencies in relation to other frequencies, and of how our preferences influence our subwoofer selections and our subwoofer placements. As the Guide has continued to expand, I have also decided to try to address some fundamental issues of speaker placement and of how rooms influence the sound we hear. And, I have added some general suggestions on techniques to use during the Audyssey calibration process.
* Much of the information in the Guide may be helpful in understanding important audio and set-up issues, and will also be somewhat applicable to non-Audyssey systems of audio/HT calibration and automated room EQ.
It may be worth pointing out that we all like having some reassurance that we are operating our audio systems correctly, with the "correct" settings, and that we are getting the maximum benefit from them. I believe though, that the more that we understand some of the basic audio principles involved (which I certainly didn't when I first got into home theater) the more confidence we will be able to have in our own individual setting preferences, and in the resulting sound quality. As with almost everything in audio, sound quality can be very subjective, and it would be very difficult to identify a single set of "correct" settings which would please everyone.
Part of the key to developing a satisfactory audio system, in my opinion, is informed experimentation. AVS gives all of us an opportunity to share information with each other, so that we can enjoy our audio/HT systems more, and be more confident in the choices we make. And, that's really what the Guide is about--sharing information and, in some cases, speculation. I have learned a lot from writing it, and continue to do so as I try to add more detailed explanations. I hope that others will find it of benefit to them as well.
Sections I Through VIII:
There are eight major sections in the Guide, which begins in this post, and continues in Post 2. All but one of the eight sections are divided into multiple subsections, which cover a wide range of related material.
I. The first section starts with a description of the frequency range that we would be discussing in our home theater (HT) systems. Following that is an extensive subsection on system setup, and how rooms influence the sound quality we hear. It offers some advice on speaker placement and some fairly detailed discussion of room treatments. It also offers some calibration technique tips that may help people to achieve better results from an HT calibration.
II. The second section explains how Audyssey calibrates our audio systems. It is broken down into subsections which are labeled. The section explains the basic principles of how Audyssey works during the set-up process, and how it EQ's our audio systems. Many of the principles explained in Section II may also pertain to other systems of HT calibration and automated room EQ.
The second section also explains how audio systems are calibrated to a Dolby/THX Reference standard. It offers some best practice advice for getting the most from our subwoofers, and explains the relationships among subwoofer gain, AVR trim levels, and master volume levels. The section emphasizes the general desirability of keeping subwoofer trim levels in the negative range and using subwoofer gain to add sub boosts. And, it explains different ways to do that.
III. Since bass management is such an important component of all our audio systems, the third section explains some basic principles to consider in setting crossovers. The LFE channel, and something called bass localization, are discussed in some detail. And, a concept called Cascading Crossovers is introduced.
IV. The fourth section explains how Audyssey, and other systems of auto EQ, calibrate and EQ multiple subs. It also explains some of the difficulties that may occur when dissimilar subs are combined in an HT system. Phase cancellation which may occur between speakers and subwoofers, and which may also occur between subwoofers themselves, is discussed in this section.
V. The fifth section examines Audyssey's DynamicEQ (DEQ) and Dynamic Volume in some detail, and also compares and contrasts Audyssey Reference and Audyssey Flat. This section also discusses the use of bass and treble tone controls, and the development of Harman and more personalized house curves. That Section V-B has general applicability well beyond the use of Audyssey.
VI. The sixth section is a brief one that explains something of the Audyssey Thread history with respect to setting subwoofer trim levels, as the current advice is different from the advice in the much older Audyssey FAQ.
The last two sections have relatively little to do with Audyssey directly, or with room EQ in general, although there are some overlaps with room EQ. But, understanding some fundamental audio concepts, and especially some bass and subwoofer concepts, can enhance our ability to get the most from our audio systems.
VII. The seventh section is a longer one which explores the way that bass frequencies behave in a room, and which explores some of the general relationships among bass frequencies,including: tactile response, how room gain amplifies our bass, and how the way we hear and feel bass frequencies may influence the settings we use. Understanding those interrelationships is important! In that section, the Equal Loudness Contours, which illustrate how human hearing works, are also discussed in detail. Ideally, Sections VII and VIII will be read in conjunction.
VIII. The eighth and final section provides some fairly detailed guidance for people who are in the process of selecting subwoofers, and also provides some basic advice on positioning them within a room. Many people start threads on which subwoofer to buy, without having a good idea of what they are actually looking for, or how to distinguish among the options which people suggest to them. Section VIII will help with that. The section starts with some general rules to follow in selecting subwoofers--sort of like the Cliff Notes at the top of this page.
The five subsections in Section VIII go into considerable detail in describing differences between sealed and ported subwoofers; some different ways to compare subwoofer performance; the pros and cons of initially buying a single large sub, versus two smaller ones; and some descriptions and comparisons of some of the more popular ID (Internet Direct) subwoofer companies. Since subwoofer placement in the room is so important, a separate subsection is devoted to that.
[It is worth noting that the Audyssey FAQ, which is linked in my signature, and especially the Technical Addendum to the FAQ, have a wealth of additional information and explanation on some aspects of Audyssey which are not covered in this Guide. Interested readers are highly encouraged to read the FAQ, for both quick answers, and for some additional in-depth detail about Audyssey. However, wherever the Guide conflicts with the FAQ, the Guide presents more current and more accurate information, as explained in the Audyssey Thread History in Section VI.]
* REW: HT owners who are encountering specific problems with their frequency responses, or who wish to optimize their frequency responses (especially with multiple subwoofers), or who are simply curious about what is actually happening in their rooms, may wish to implement REW, which is a free download. Sometimes, people get very acceptable results from their automated room calibration systems, and sometimes the results are not satisfactory to them. In my opinion, this is an entirely personal decision which individual listeners will make for themselves.
In any event, measuring their frequency responses can tell those who are interested a lot about proper subwoofer positioning, set-up, and post-calibration adjustments. The use of REW will require a calibrated measurement microphone (a UMIK-1) and a computer (preferably a laptop) which can be connected to their AVR's or AVP's. Anyone interested in learning more about REW, and how to implement it, is encouraged to consult the following step-by-step guide by AVS member @AustinJerry.
https://www.dropbox.com/s/zdhq72a1puyyxpr/REW 101 HTS Current Version.pdf
There is also an AVS discussion thread which concentrates on the practical application of REW:
Guide to Subwoofer Calibration and Bass Preferences
Section I: Room/System Setup and Sound Quality
There are a number of factors which can affect the sound quality in our rooms. Those factors include our speaker choices and their placement, distortion from the room itself, and the use of room treatments and automated room correction. There are a number of potential reference sources which can help us with our initial system setups, in terms of positioning our speakers, or with room treatments, or with specific room EQ calibration tips. But, I think that it would be worthwhile to try to address some basic concepts in this Guide, so that it can be a more general resource.
With that in mind, I would like to try to explain some basic concepts of system/room interaction and to offer some general advice on the relationships among the room, our system setups, and our sound quality. I would also like to offer some general tips on performing a successful calibration with room EQ.
It would probably be helpful to define some terms that are used in audio, and throughout the Guide. (Additional definitions and abbreviations are presented as they are used in individual sections.) I will start by using a good online definition of sound. Sound in air is made when air molecules vibrate, and move away from the vibrating source, in a pattern we refer to as sound waves. In our context, the vibrating source would be our transducers--our speakers and our subwoofers.
Sound pressure level (SPL) is a measurable quantity of sound volume. It is measured in decibels (dB). "Loudness" is not a measured quantity of volume; it is a perceived amount of sound. For instance, "That sounds really loud!" is a very different statement than "The SPL in the room is 100dB, as measured at the main listening position (MLP)." Loudness is a perception of how something sounds, while SPL is a measurable quantity of sound volume. The distinction between those two terms becomes very important when we are selecting our preferred listening volumes and our bass volumes.
The Subsections in Section I are as follows:
A: The Frequency Range
B: Distortion, Speaker Placement, and Room Treatments
C: Room EQ and Calibration Techniques
Section I-A: The Frequency Range:
Since we will be talking about various frequencies and how we hear them, throughout the Guide, I think that it would be helpful to begin with some explanation of how I would personally subdivide the frequencies that might be part of an audio/HT discussion. The lowest bass that can be meaningfully reproduced in an HT system is approximately 7Hz, although it is unlikely that most of us would be able to hear that low with complex content. Very high-frequencies can potentially be reproduced by modern tweeters, but the absolute upper limit of young healthy human hearing is 22,000Hz.
So, that 7Hz to 22KHz range is the one that we will be focusing on. But, how do we subdivide that range into divisions that facilitate a discussion of speakers, subwoofers, room treatments, and all of the other HT-related subjects that we may be interested in? That is the purpose of this subsection. I will begin this subsection by sharing two completely different graphs of the frequency range that we are discussing. We could find many others with a Google search.
The second graph is a little more complete than the first one, so in the discussion that follows, I will make some references to the Harman graph just above. It charts an actual in-room measurement of frequencies, with a rising-bass house curve added. (House curves are described in Section V-B. But, to briefly synopsize, we don't hear bass frequencies as well as other frequencies, so most people prefer to increase bass volumes, relative to those in our normal hearing range.)
I find discussions of bass frequencies very interesting, and also very confusing. I have researched this topic on numerous occasions, and have always found completely different ways to subdivide the frequency range of human hearing. There is particular disagreement as to the upper limit of what is a "bass" frequency, as opposed to a mid-range frequency. (There is also disagreement about what exactly is a mid-range frequency.)
I think that if we look for graphs online, we will find some agreement that ULF (ultra low-frequency) bass is <20Hz, although many frequency graphs stop at 20Hz. And, we will frequently see the range between ULF and 50Hz defined as the low-bass range. (FWIW, I think that 30Hz also has some special significance, as that is where we start to have trouble distinguishing between sound and physical vibrations.) Where mid-bass is defined at all, it will typically start at 50Hz (although it's sometimes 60Hz) and it often extends up to about 100Hz. (I personally prefer to use 120Hz, for reasons that are explained below.) But, that still leaves the upper-bass range, and that dividing line is all over the place.
Here is part of the problem as I see it. Everyone who is attempting to define bass frequencies is approaching the definition from a slightly (or dramatically) different perspective. Some of the people who are defining bass, mid-range and high-frequencies are musicians, and they tend to approach the issue from the standpoint of the musical instruments themselves. For instance, a 4-string upright bass has fundamental frequencies that range from a low of 40Hz to a high of 400Hz. So, since the upright bass is specifically designed to be a bass (and a deep-bass) instrument, some musicians might define bass frequencies as extending to about 400Hz.
Alternatively, some of the people who offer subdivisions of bass frequencies are recording mixers. And, by and large, I think that most of them define bass frequencies as extending to somewhere between 200Hz and 300Hz. It is shown as 250Hz in the Harman graph. In some respects, that 200Hz to 300Hz definition of the upper limit of bass frequencies seems even more arbitrary than the musical one, with each graph using a different dividing line.
Still another definition of bass frequencies comes from some audio engineers and HT hobbyists, who think in terms of the Schroeder (transition) frequency in a room. That is the frequency where low-frequencies become standing waves inside a room. Most of those definitions put an upper limit on bass frequencies of about 200Hz, because that seems to be about the upper limit of that transition frequency even in very small rooms.
I say that it "seems" to be the upper limit, because if you attempt to use any of the online calculators to determine the transition frequency in a room, you will get widely divergent results from the different calculators. If our definition of what is an upper limit of bass frequencies is dependent on the room size and construction, the specific room geometry, and the reverberation time within the room, then this may actually be the least useful definition of all. The definition of what is the upper limit of bass frequencies will vary with every room under discussion.
Another definition of the upper limit of what is a bass frequency comes to us from three-way speaker designers, who design their woofers to have upper limits, and their mid-range drivers to have lower limits, and who then create internal crossovers between the two. Those crossovers between woofers (bass drivers) and mid-range drivers is typically somewhere between about 300Hz and 400Hz, although some three-way speakers are probably outside of that range on either end. (It is also important to note that a woofer in a three-way speaker has to be able to play a little above the crossover. So, if the crossover is at 300-400Hz, the woofer has to be able to play frequencies up to at least 500Hz and higher.)
Here is an example of a definition that attempts to bridge several approaches. This one comes from a speaker designer and the bold emphasis is his. I will let individual readers make their own sense of this one:
"The bass frequencies cover 20 to 1,000 hertz, while treble covers 1,000 to 20,000 hertz and mid-range overlaps from 300 to 3,000 hertz. Mid-bass range is approximately 140 to 400 hertz. A mid-bass woofer is a speaker specifically designed to handle this sound frequency."
I will give that one points for originality, if for nothing else. Confused yet? I certainly am! Given the wide disparity in definitions, and the lack of an apparent logical basis for most of them, I decided several years ago to come-up with my own divisions for use in the Guide. At least that way I can explain my reasons for selecting the division of frequencies that I use. The divisions I use have evolved a bit from where I originally started them, as I have learned more, and thought through things a little differently. I have also wanted to achieve better consistency in my methodology, and I believe that my current approach does that.
I don't claim that my division of frequencies is "correct" in any universal sense, or even in any specific use of that term. It is simply a reasonably logical way of dividing the frequency range, that I think may be useful in discussing audio, and its application to our home theaters.
Dividing the Frequency Range:
Starting with bass, I would define the upper limit of the bass range in the following way. To me, it is approximately 500Hz (~480Hz according to the Equal Loudness Contours), where our perception of loudness starts to change. Below that frequency, we require more volume to hear sounds in equilibrium with those in our normal hearing range. That seems to me like a logical place to say that bass frequencies are starting. And, it's only a little higher than a couple of the other definitions that we saw.
Supporting that upper limit is the fact that frequencies below about that frequency begin to radiate more omnidirectionally, rather than in a more directional fashion. That means that the bass frequencies are leaving a speaker cabinet in all directions, rather than just coming from the general direction of the speaker cone. As noted, most speaker makers also cross from mid-range drivers to woofers in about the 300-400Hz range, although some cross a little lower or higher than that. So, if we established 500Hz as the upper limit of bass frequencies, I think that we would be in the right general ballpark for most HT discussions.
(I could also support a division for bass that was a little lower than 500Hz; perhaps in the 300-400Hz range. But, for the purposes of the Guide, I prefer the logical consistency of using the Equal Loudness Contours, and the way our hearing changes at about 500Hz, for this division.)
Mid-bass frequencies may also be a little difficult to define, and we don't always see that range specified in frequency graphs, such as in the two examples above. But, in HT, it seems to be a pretty important range, which comes-up all the time when people are selecting, configuring, and EQing subwoofers. I like defining mid-bass as the range from about 50Hz to 120Hz. That's a fairly small range, but it has several things to recommend it. First, for most people, that seems to be the average range where chest punch sensations are felt, and most people already associate those chest punch sensations with mid-bass frequencies.
I emphasize the word "most" here, because our perception of chest punch probably follows a bell curve, with some people outside the norm at both ends of the curve. Several studies have reached similar conclusions on that approximate 50Hz to 120Hz range, with one blind study determining that 63Hz was the frequency where most of the participants felt chest punch most strongly. At least two ID sub makers provide a pre-programmed PEQ boost centered on that specific 63Hz frequency.
To me, another reason that the 50Hz lower limit makes some sense for mid-bass, is because that is about the frequency where we can often observe a difference in the performance of ported subs and sealed subs, where sealed subs are starting to roll-off compared to ported subs, which maintain linearity at that frequency. There can be exceptions to that generalization, especially with the very largest and most powerful sealed subs. But FWIW, I like defining frequency ranges that correspond to some pragmatic HT considerations.
Using 120Hz as the upper end of the mid-bass range also makes some sense to me, since that is the upper limit of the .1 low-frequency effects channel (LFE) which was specifically intended to be played by subwoofers in the original Dolby/THX standards. The low-pass filter for the LFE channel is not a brick wall, and some sounds creep-in above 120Hz. But, sound mixers are primarily trying to amplify specific bass content in that channel only up to 120Hz. (Or perhaps, only up to 80Hz with respect to the most meaningful bass content.) In any event, that 120Hz low-frequency effects cutoff point seems like a logical separation between mid-bass and upper-bass.
There can also be multiple ways to define what constitutes low-bass. For instance, as noted earlier, many graphs which describe the frequency range don't even consider frequencies under 20Hz. So, low-bass would just be anything under about 50-60Hz. If we define low-bass as about an octave-and-a-half below the 50Hz limit that we set for mid-bass, and ULF as <20Hz, we have the following relatively proportional divisions, which go from the lowest frequencies to the highest frequencies:
* 7Hz to 20Hz: ULF, which covers the frequencies below 20Hz. That would be about 1 1/2 potentially meaningful octaves, using the 8-note per octave scale, where each doubling of frequency is one octave. (For example, the frequencies between 10,000Hz and 20,000Hz would still just be one octave, consisting of 8 distinct notes.)
* 20Hz to 50Hz: Low-bass would be about the 1 1/2 octave range from 20Hz to 50Hz.
* 50Hz to 120Hz: Mid-bass would be the roughly 1 1/2 octave range from 50Hz to 120Hz (125Hz would make it an exact octave-and-a-half, but that's splitting hairs.)
* 120Hz to 500Hz: Upper-bass would be the 2-octave range from 120Hz to approximately 500Hz. Below 500Hz is where our perception of equal loudness starts to change.
* 500Hz to 5,000Hz: I would define the mid-range frequencies as covering the frequency range from about 500Hz to 5,000Hz. That is just a little more than 3 octaves, which seems about right for the frequency range where our hearing is the strongest. Most speaker designers seem to cross their mid-range drivers to their tweeters at just about 2,500Hz to 3,000Hz. In fact, mid-range compensation, found in some audio curves such as Audyssey's default Reference curve, is centered on 2,500Hz. That -3dB reduction in SPL at 2,500Hz is based on the original "BBC dip", which was designed to improve crossover blending from mid-range drivers to tweeters.
Since mid-range drivers need to play a little above a crossover of around 2,500-3,000Hz, it makes some sense to define the upper limit of the mid-range as 5,000Hz, in order to provide some cushion for that. Part of the reason for using 5,000Hz, as a dividing line for mid and high-frequencies, is also for the sake of consistency with what we used for the upper limit of bass. Our normal hearing range, where all frequencies sound equal in loudness, is the range from 500Hz to 5,000Hz. According to the Equal Loudness Contours, our perception of loudness changes at 5,000Hz, with frequencies above that sounding a little softer, just as they start to do below 500Hz. So, it makes a certain amount of sense that the "middle range" would correspond to our normal hearing range of 500Hz to 5,000Hz.
* 5,000Hz to 22,000Hz: Treble or high-frequencies. Based on my current thinking, high-frequencies would start at about 5,000Hz, and continue all the way up to 22KHz, which is the extreme upper limit of young and healthy human hearing. (Most of the people reading this, including the person writing it, probably can't hear much above about 12KHz anymore, if we can hear even that high. But, that's another story.) That frequency range would be approximately 2 octaves.
Most music-related definitions of high-frequencies draw a distinction between fundamental frequencies, which only extend up to about 6,000Hz for almost all musical instruments, and harmonics (one and two octave overtones) of those frequencies, which add 'brilliance' to the sound. There is a graphic illustration of the range of musical instruments in Section I-B. They then generally subdivide the mid-range category, extending up to about 4,000Hz, into three separate divisions: low-mid, mid-mid, and high-mid, as illustrated in the Harman graph.
Between 4,000Hz and 6,000Hz, in that same Harman graph, is something called "Presence". Everything above 6,000Hz is then considered "Brilliance". I don't personally find any of the low-mid, mid-mid, and upper-mid divisions to be particularly useful for audio/speaker/HT purposes. Nor, do I find the terms "Presence" and "Brilliance" especially helpful for HT use.
And, since some musical instruments do play fundamental frequencies above 6,000Hz, that first division at 4,000Hz, and the second one at 6,000Hz, seem somewhat arbitrary to me. I have some idea of what is meant by "presence" in musical terms, but the term doesn't carry enough intrinsic meaning to be very helpful in our HT discussions. I do think that the use of the term "brilliance" can be helpful, in a general descriptive sense, for musical instruments. And, I have also sometimes used the term "bright" or "brilliant" to describe the high-frequency sound of some tweeters, or of high-frequencies inside a relatively untreated room. But, for HT discussions, at least, it may be a little too ambiguous a term to constitute a meaningful frequency division.
(FWIW, I think that the difference between 5,000Hz and 6,000Hz is pretty inconsequential when we realize that there are only 8 distinct notes in the octave between 5,000Hz and 10,000Hz. So, if someone else wanted to define high-frequencies as starting at about 4,000Hz, or at 6,000Hz, I certainly wouldn't have a problem with that. I would hope though, that there would be some specified basis for the definition, as there is here.)
In any event, I like the fact that this overall subdivision of frequencies seems to have some inherent logic and proportionality. And, for audio/HT purposes, I find that definitions of frequency ranges are more helpful, when they somewhat align with the way that our speakers and subwoofers work, and with the way that we actually hear different frequencies. To me, if we just think of the way that a three-way speaker works, we gain some insight into how to subdivide frequencies for HT discussion purposes. Woofers play bass frequencies (with some overlap above their internal crossovers), the mid-range drivers play mid-range frequencies (with some overlap both above and below a crossover), and the tweeters start playing softly below the crossover to the mid-range driver, and then play all of the treble frequencies.
Once again though, I think this points to the real nature of the problem in these discussions. Are we coming at our divisions of the frequency range from the standpoint of the operating range of musical instruments, or from the perspective of recording mixers, or are we thinking of subwoofer/speaker performance in an HT? For HT purposes, I find the subwoofer/speaker perspective (combined with the way that our hearing actually works) to be the most useful way to talk about divisions within the frequency range of about 7Hz to 22KHz. But as with almost any audio/HT-related issue, there can be other legitimate viewpoints.
Section I-B: Distortion, Speaker Placement, and Room Treatments
What we hear, when we listen to a recording, is heavily influenced by the room itself. To begin the discussion in this section, it may be helpful to talk a little about some factors that influence what we hear, and to talk about room-related distortion in a general way. There are much more sophisticated explanations of distortion that readers can investigate, but I want to try to cover some very basic concepts. First, distortion can be defined as "any alteration in the waveform of an audio signal." (All sound is composed of frequencies--sound waves--so any alteration in the waveform can affect what we hear.)
In practice, "distortion" is sort of a catchall term for nearly anything which adversely affects perceived sound quality. Some audio sources we listen to may have compression in the source itself, or distortion in the specific recording. Speaker, subwoofer, and AVR makers go to great lengths to minimize distortion, and particularly Total Harmonic Distortion (THD) in their products. Anyone interested in understanding more about audio distortion, from audio sources and transducers, is encouraged to consult more authoritative articles such as this one:
Blog - How Much Distortion Can We Hear With Music? | Axiom Audio
Since frequency response (FR) is mentioned so often in audio discussions, and in the Guide, it is also worthwhile to define what is meant by a "smooth", or "even", or "flat" frequency response. A smooth, even, or flat frequency response would be one in which no frequencies are playing significantly louder or softer than other frequencies. (We will often refer to peaks and dips in the FR.) In music, and in movies, some sounds will be deliberately emphasized over other sounds--such as a trumpet solo in a music recording, for instance, or low-bass sequences in movies. That would be inherent to the recording and not dictated by the room. However, the room will inevitably influence the evenness of the FR in unpredictable ways, causing some frequencies to sound louder or softer than others. We can remediate that potentially uneven FR with good placement of our transducers, with room treatments, and with room EQ.
Some of us may be deliberately creating a house curve to change the interaction of our recordings and our room, in some specific direction (such as having less treble, or having more bass). But, most of us would not typically want the room itself to arbitrarily dictate which frequencies are louder or softer than others, in a way that has nothing to do with the original intent of a recording. Where a frequency response is reasonably "flat" or "smooth", all frequencies are playing at approximately the same volume, at a particular point in space, unless we, or the recording, dictate otherwise. The selected point in space where everything is more-or-less in equilibrium is the main listening position (MLP). As we move away from the MLP, the relative smoothness of the FR may change, as some frequencies become slightly louder or softer, compared to others.
1. Room Distortion:
If we drive a particular speaker, or subwoofer, or AVR too hard, we may introduce distortion into our signal chain. But, even if we are playing content at moderate volumes, once we play an audio recording, from any kind of transducer (a speaker or subwoofer which converts electrical energy into sound) located inside a listening room, the interaction with the room itself influences the sound that we hear. Not all influences are negative, as the room may influence the sound we hear in a number of positive ways. But, not all room influences are positive, either. It is the negative influences caused by interaction with a room, which we may characterize as distortion, that I am going to try to address in this brief discussion.
Section VII explains in more detail the difference between frequencies below the Schroeder, or transition frequency in a room, and frequencies above that point. Briefly though, below the transition point in a room, which is typically about 200Hz or lower (depending on the size of the room), some bass sound waves slow down when they encounter a room boundary, and they collect as "standing waves". Some bass frequencies are amplified, and some are cancelled, as a result of the interaction of those standing waves with room boundaries. Room boundaries would typically consist of four walls, the ceiling, and the floor, although there could be more or fewer than four walls in a given room.
Overall, room boundaries reinforce the bass SPL, but the amplification/cancellation of individual frequencies can be problematical. And, they are inevitable whenever a bass transducer is placed inside a room. (We can mitigate that somewhat with good speaker/subwoofer placement, and with the proper placement and integration of multiple subwoofers.) Higher frequencies, especially above about 300Hz, behave differently than the lower frequencies do. They aren't amplified or cancelled in the way that low-frequencies are. But, when direct sounds and reflected sounds from those higher frequencies arrive too closely together at a listening position, they can cause distortion in what we hear.
Mid-range and treble frequencies may not really reinforce the measurable SPL in a room, but they certainly may sound as if they do, especially in very lively-sounding rooms. In a room with a lot of hard surfaces, some of what we are perceiving as loudness is probably distortion. (I am using distortion in this case as a catch-all term. In a small room, sound waves bounce back-and-forth for a longer period of time than they would in a much larger room. The more bare surfaces there are in the room, the more prolonged reverberation we will have.) It is easy to understand this if we think of the difference between singing in a shower, with bare walls and a tiny space for the sound to reverberate in, and singing at the same volume level in a normal size bedroom.
We typically perceive high-frequency distortion as being louder than undistorted sound. In fact, we may perceive it as being much louder. And, most of us don't seem to tolerate higher frequency distortion as well as we do lower frequency distortion. In fact, for some people, high-frequency distortion can be a little painful. The term "ear fatigue" is sometimes used to describe hearing high-frequency distortion during a listening session.
The words we use to describe different types of distortion tell us something about our discomfort. For instance, we may describe high-frequency distortion as screechy or piercing (which suggests a somewhat painful sound), where we might describe low-frequency distortion as muddy or boomy (simply suggesting a lack of clarity). The words we use to describe the different types of distortion are significant, and the difference in the way we hear distortion at different frequencies can be important. Understanding differences in the way we hear some frequencies, and learning something about the distortion that may accompany those frequencies, can be helpful as we try to improve the sound quality in our rooms.
[It may be important to understand that home theater (HT) rooms of about 20,000^3 (^3 is an abbreviation for cubic feet) or smaller, sound perceptually louder than commercial cinemas, which are always much larger than even very big HT's. A number of audio experts estimate that our HT's may sound anywhere from about +5 to +9dB louder than the same SPL would seem when played in a commercial cinema. The +5dB to +9dB range of difference varies with the size of the room, with very small rooms (<1500^3) being as much as +9dB louder sounding than the same volume would be if played in a commercial movie theater. The following table illustrates the relationship between room size and perceived loudness, based on an assumed volume level of 85dB in a commercial theater. As you can see, in a room smaller than about 1,500^3, only about 76dB would be required to equal the apparent loudness of a commercial theater playing at 85dB:
As noted earlier, some of that louder-seeming sound might potentially be due to distortion. But, even in a heavily treated room, the smaller size of the room would appear to amplify the loudness level we would perceive. I have heard room treatment consultants estimate that a well-treated room might sound up to about -3dB less loud, than it did prior to room treatment. I suspect that the degree of loudness attenuation might depend somewhat on the size of the room. For instance, a very small room might benefit even more than -3dB from extensive room treatments. But, it would still sound significantly louder than the same SPL played in a commercial theater.]
In order to understand how rooms can create distortion, it is important to understand something about how sound waves behave in a room. For the purpose of the discussions in the Guide, I am defining bass frequencies as those frequencies below 500Hz. As noted previously, I would probably define mid-range frequencies as the range from 500Hz to about 5,000Hz. As a practical matter, frequencies from about 2,500Hz or 3,000Hz and up would typically be played by tweeters, as they have to play slightly below the internal crossover in a speaker.
This may be a good place to illustrate a description of what musical instruments play what frequencies. Understanding that will help to correlate what is written in this Guide, and elsewhere, with what we are hearing with acoustic (non-electronically enhanced) instruments. The following graphic is one popular interpretation of the frequency range of musical instruments. The red horizontal lines represent fundamental frequencies, and the yellow lines represent harmonics (overtones) of those frequencies, which add brilliance or sparkle to the fundamental sounds that we hear.
Sound waves are vibrations, with different lengths and different vibration speeds. (They vibrate--moving back-and-forth, sort of like a slinky toy, as they travel through the air.) They all travel at the speed of sound (which is approximately 1 millisecond per foot at sea level), but how they behave in a room, and how we perceive them when they reach us, makes all the difference in what we hear.
As frequencies get higher, the vibrations become much shorter in length, and they oscillate (move or travel back and forth between two points) much faster. They also tend to travel in straight lines unless absorbed, or redirected, by contact with some surface. Longer frequencies oscillate more slowly, and they tend to bend rather than always simply ricocheting in a straight line. They also go right through solid objects in a way that higher frequencies cannot. Those differences can have a direct bearing on what we hear in a room. And, the differences in sound wave lengths are very significant.
For example, a 10Hz frequency is only 10 cycles per second (meaning that it vibrates 10 times per second), and the wavelength is about 112' long. By contrast, a 100Hz frequency is 100 cycles per second, and the wavelength is now only about 11' long. As we go up in frequency, the wavelengths get shorter and shorter. A 1,000Hz frequency is 1,000 cycles per second, and the wavelength is just a fraction over a foot long. By the time we are up to 10KHz, and 10,000 cycles per second, the wavelength is only a little more than an inch long. The table linked below illustrates frequency length:
Frequency - Wavelength - Period Chart
Where low-frequencies oscillate relatively slowly, bending and pooling in corners (where any two or more room boundaries meet), the frequencies above the transition point (which is usually above about 200Hz or so in most rooms) oscillate increasingly faster, and bounce around in a room, reflecting off of hard surfaces like a billiard ball bouncing off the cushions on a billiard table, until they run out of energy from friction. The billiard ball analogy actually falls a little short, as all frequencies, but especially mid and high-frequency sound waves, ricochet off of all six surfaces in a room (the four walls, the floor, and the ceiling) until they run out of energy, or are absorbed by something. (An even better analogy would be a handball court.) Higher-frequencies are absorbed more easily than low-frequencies, and they run out of energy faster due to heat exchange with the air and with the surfaces they touch, than is the case with low-frequencies.
Sound waves may also bounce off of other hard surfaces, such as table tops or other furniture. Those reflected sound waves which don't arrive very close in time to the first sounds to reach our ears, may be somewhat ignored by our brains. We typically hear the first-arriving sounds as being louder, and we concentrate more on those sounds. Our brains can typically filter-out, or separate, the later arriving sounds, if the sounds are somewhat delayed, in order to concentrate on those louder, first-arriving sounds. It's sort of like unconsciously tuning-out the conversation at an adjoining table, in a restaurant, if we are interested in what our dinner companion is saying.
But, higher frequency sounds arriving within about 6 milliseconds (ms) or so of each other are often associated with distortion, because our brains can't distinguish them as clearly from the first arriving sounds. (As noted earlier, sound travels at approximately 1ms per foot at sea level, and very slightly slower than that at high altitudes due to cooler temperatures.) I suspect that how much we might notice distortion, from sounds arriving too close together, can vary somewhat depending on the individual, and on both the frequency and the SPL of the sounds. For instance, distortion usually becomes more noticeable at higher volume levels.
In any event, sounds arriving close behind the first-arriving sounds from our speakers, within approximately that 6ms window or so, may distort the sound we hear. In a best case scenario, early reflections may simply make some sounds seem louder or more three-dimensional. But, they may also be perceived as contributing to a harsher sound with some mid-range frequencies or to a more strident or piercing sound with high-frequencies.
To summarize, a reflected mid-range or high-frequency sound from a wall or other hard surface, which arrives at our ears at almost the same time as the direct sound from a speaker, may create a type of distortion, as our brains will have difficulty combining the separate close-arriving sounds into a single coherent sound. When that happens, we may experience ear fatigue, or we may simply suffer a loss of clarity in the sound we hear.
Distortion may, or may not, be particularly noticeable, depending on whether we are actually accustomed to hearing undistorted sound from our audio systems. This is an important point! The expression that "we don't know what we don't know" could be expanded to "we may not hear what we don't know to listen for". A personal example of that is given later, involving employees at a high-end audio store who were always accustomed to hearing distortion in their store's listening room, and who consequently didn't know that the music recordings they played for customers in order to audition expensive speakers and amps weren't really supposed to sound that way.
And, there can also be a component of individuality involved. Not everyone is going to hear exactly the same frequencies, in exactly the same way, and our listening preferences may vary as well. As volumes increase however, excessive early reflections may cause distortion which is painfully shrill or harsh to some people. As noted, "ear fatigue" or "listening fatigue" are terms which may often relate to some form of mid-range or high-frequency distortion.
Our initial system setup can contribute to the relative clarity and fidelity of the sound that we hear in a room. Some suggestions regarding subwoofer placement in a room are addressed in the last section of the Guide. In this first section, though, I would like to concentrate on the speakers, and on their relationship to the room and to the listening positions. I would particularly like to concentrate on the speakers located along the front soundstage, as they carry the majority of the content in a movie or music recording, and will have a much greater influence on our overall sound quality.
(There are numerous on-line guides illustrating placement objectives for surround and overhead speakers, but I believe that insufficient attention may often be paid to the front soundstage which is so critical for both music listening and movie/TV viewing.)
In the context of the discussion of how sound behaves in a room, anyone who is interested in how specific frequencies sound, and how the room may influence our perception of sound, may find the following video helpful. In the video, individual frequencies are played, starting at 20Hz and continuing to 20,000Hz. The test should be conducted only in stereo, and the sound should always be coming from the center. Where the sound seems to come more strongly from one ear than from the other, the room your speakers are located in is influencing the sound, as room reflections from one speaker or the other momentarily seem to dominate the sound. It's a great illustration of how the room influences what we hear.
Something else that the video demonstrates is how low some low-bass frequencies may actually sound to us. Most of us probably believe that some frequencies we hear are lower than they really are. Getting a feel for how low in frequency an 80Hz, or 60Hz, or 40Hz tone actually is can be quite revealing. That is why a consistent theme in the later sections on bass is that we feel as much as hear frequencies below about 30Hz. And, it can take a fair bit of SPL to hear and feel those very low-frequencies, in conjunction with the sounds in our more normal hearing range, which may be accompanying them.
Pure tones, like those in the video, and the complex sounds which we hear in music and in movies, are very different. Complex sounds may consist of multiple frequencies, each containing both fundamental frequencies and harmonics (overtones) of those frequencies. Most of the time, the very low-frequencies are simply adding some bass weight to sounds, rather than standing-out as distinct sounds themselves, although there can be some exceptions to this with some low-bass content.
2. Front Speakers:
To continue the discussion of room distortion, let's start with the front speakers. Many speakers require some distance from the wall behind them in order to achieve good sound quality. Electrostatic speakers may require even more separation from a wall as they don't radiate sound in the same fashion as direct-firing speakers. Bipole speakers are designed to radiate sound both forwards and backwards toward walls. And dipole speakers are designed to radiate most of the sound backwards toward walls, in order to create a more diffuse sound field.
Most speakers are direct firing--meaning that the sound primarily travels in the direction they point toward. With direct-firing speakers, the sound leaves the drivers, in the front of the speaker, in a cone-shaped cluster of sound waves. But, even with direct-firing speakers, other sound waves travel backwards, through the speaker cabinet, to reflect off the wall behind the speakers. (Rear-ported speakers may also require some wall clearance to operate properly. About 4" is a good rule-of-thumb for port clearance.) Bass frequencies will radiate omnidirectionally, from all sides of the cabinet.
With front speakers something to be aware of is called SBIR. It stands for Speaker Boundary Interference Response. It may or may not ever be an issue to specific listeners, but it is worth mentioning. If a speaker is too close to a wall, there may be some boundary interference from that wall. Low-frequencies radiating backwards, or toward a sidewall, would strike the wall and reach the listening position less directly then the waves traveling in a straight line from the speaker. That, in turn, could cause them to be out-of-phase with the direct wavelengths.
Where that happens, some bass frequencies would be amplified, while others were cancelled, causing peaks and dips in the overall frequency response. In some cases, the dips will be significant. But, in some respects, the peaks will be even worse. Our brains will concentrate on first arriving sounds and will to some extent disregard later arriving sounds. But, if direct and indirect sounds arrive too close together, they will create distortion in the sound. Improving the effects of SBIR can not only allow us to hear frequencies that might previously have been cancelled, but it will improve the overall clarity of the sound.
Moving front speakers away from walls can help to prevent SBIR. Placing approximately 2" to 4" thick broadband acoustic panels behind speakers which are close to walls, can also help to prevent boundary interference, where it is believed to be a problem. A good way to determine whether or not SBIR is a problem is simply to experiment with moving speakers closer to, and further from walls, and listening to the results. You are simply listening for the placement where the sound is the clearest, prior to running room correction.
The following video is excellent in explaining what SBIR is and how to deal with it:
To return to the general discussion of speakers, several things are happening when a speaker plays inside an enclosed space, such as an audio/HT room. First, the direct sound is arriving at the listening position. Second, sound waves which bounced off the wall behind the speaker, are arriving just a little later at the listening position. Third, sound waves which bounce off the side-walls are arriving at the listening position just a little behind the direct sound. And fourth, sound waves which bounce off the floor and ceiling (and to some extent off the wall behind the listener) are arriving at the listening position just a little behind the direct sound. And, all of those sounds will arrive very close together. Our brains are remarkably adept at sorting-out all of those very early and later-arriving sound waves into a single coherent sound. But, when the sounds arrive too close together, they can't, and we hear distortion.
Opinions vary as to how to address distortion caused by the speakers and the room. For instance, speaker makers typically try to walk a fine line with respect to having drivers with wide horizontal dispersion, in order to provide a wider-sounding soundstage and in order to maintain the same SPL, off-axis, as they do when they are pointed straight ahead. But, if they get too wide, early reflections from side-walls may become more of a problem. Again, it is helpful to remember that sound waves above about 500Hz leave a typical direct-firing speaker in a cone shape, with the narrow portion of the cone closest to the speaker, and the mouth of the cone getting wider as the sound waves move further away from the speaker.
I think that most speaker makers try to achieve about 30 degrees of horizontal off-axis dispersion, at a distance of about 3 to 4 meters. That would represent the typical listening distance for most tower speakers, or for larger bookshelf speakers. Vertical dispersion is usually more restricted. I believe that about 15 degrees is typical in order to avoid too much reflection from floors. It is important to note that some very large speakers are intended for relatively large rooms, and for longer listening distances--up to 5m or more. Knowing this can be important, because large tower speakers, which are too close to a listening position, may not be able to achieve as good inherent sound quality, without the intervention of room EQ, as speakers which are designed for that typical listening distance referred to above.
[This issue of having large speakers, designed for greater listening distances, may be worth pursuing a little further. If we think of a typical three-way tower speaker, the speakers (woofer, mid-range, and tweeter) are arranged vertically, with each driver spaced a prescribed distance from the other drivers, so that they will voice appropriately at a particular distance. (There is more to it than just the driver spacing, but this conveys the general idea.) That means that the drivers are designed to all play at the same volume at a designated distance. The distance is usually a range. As noted above, I believe that 3 to 4 meters is typical for a tower speaker, or for a large bookshelf speaker.
Smaller bookshelf speakers, however, may be designed for a listening distance of only about 2m. For instance, My two-way Audioengine A5+ speakers only have a tweeter and a 5" mid-range/woofer driver. They are designed to be used primarily as desktop computer speakers, without an external amplifier or AVR. And, they are self-powered for that purpose. They are intended to be separated by anywhere from about 3' to 6' from each other, and they are designed to voice most effectively at about those same distances. To facilitate that, the drivers in each speaker are in a relatively short vertical cabinet, and the drivers are spaced closely together. Spaced appropriately, they image extremely well in my desktop system, with good sound separation and a good phantom center in stereo.
A very large tower speaker which is designed to be employed in a large room, on the other hand, may have speakers which will play at equal volumes at a distance of 5-6 meters or more. I actually have a pair of very large (18^3) speakers like that. They were specifically designed to be separated from each other by about 20' and to each be about that same distance from the main listening position. They were always intended for larger listening venues, and they don't really sound as good in a smaller room.
Automated room correction can help the various drivers, within a vertical cabinet, to play the same volumes at the calibrated listening distance. (It will amplify some frequencies to play at the same sound level as other frequencies.) But, starting with speakers which are actually sized reasonably appropriately, for the listening distance, is still a good idea. And, this is a concept which is sometimes overlooked when people select speakers. They may select large tower speakers, which voiced well in an audition room at an audio store, but which don't sound quite as good, at their own listening distance, in a smaller room.
As noted above, automated room EQ can help somewhat to alleviate issues with our listening distance from our speakers, because EQ filters (control points) can be set at various frequencies to insure that all of the frequencies reach the listener at the same volume level. That is slightly different from the distance setting which governs the arrival time of the overall sound from a speaker.
But, when speakers are situated at approximately the appropriate distance from a listener, and at the appropriate height with respect to the listener, all of the drivers in a well-designed speaker should start-off playing at about the same inherent volume level, at least prior to room influences. That makes it much easier to achieve a smooth frequency response, and good sound quality, even before automated room EQ sets filters to improve the speakers' specific interaction with the room. It should also be noted that not everyone will have room correction to start with, or will choose to use room correction for higher frequencies. More on that later.]
Finding the sweet spot for a particular pair of speakers, in a particular room, usually requires some experimentation. For instance, if a speaker is pointed too far away from the listening position, there may be an adverse impact on the high-frequencies (which are typically very directional), and it may result in an increase in early reflections from the side-walls. So, determining how far to keep speakers away from the wall behind them, and from the wall beside them, and how much to toe-in the speakers toward the listening position, is something that may require some trial-and-error. Ideally, listeners will try to perfect the sound quality of their speakers, in their rooms, before running room EQ. That is because the better the native response is to start with, the less that room correction is likely to interfere with the quality of the sound. (There is more on this subject in the following subsection on calibration tips.)
The old stereo rule-of-thumb about trying to keep a roughly equilateral triangle with respect to the two front speakers and the MLP (main listening position) is still a pretty good starting point, in order to determine the best separation for the front speakers. That doesn't relate directly to distortion, but it does relate to the ability to easily generate good stereo and HT imaging, and it may also relate to the ability to create a realistic soundstage width, which extends beyond the screen for HT, or beyond the speakers for stereo. So, soundstage width, and the overall benign or malign influence of wall reflections, may also be factors when trying to decide how close front speakers should be to side-walls.
Another factor involving the placement of the front speakers concerns the early reflections from side-walls referred to above. It was suggested earlier that most sound waves leave a direct-firing driver in a cone shape, and that most speaker makers try to get about 30 degrees of off-axis horizontal dispersion. Depending on how close a speaker is to a side-wall, and how much (or how little) the speaker is toed-in toward the MLP, the more that early reflections from side-walls may be an issue. That might be especially true for speakers where speaker manufacturers specifically suggest pointing speakers straight ahead, or even toeing them away slightly, as that could result in more side-wall reflection.
Speaker positioning, with respect to early reflections from adjacent walls, is entirely a matter of personal preference. Some manufacturers of acoustic panels, and some audio experts and AVS members, believe quite strongly in always treating side-walls for early reflections. On the other hand, listening tests conducted by Dr. Todd Welti, Dr. Floyd Toole, and others, indicate that most people prefer the additional ambient sound provided by untreated side-walls.
According to those listening tests, most people hear a wider and more realistic sounding soundstage when those side-wall first reflection points are either untreated, or at least, less heavily absorbed. Something which disperses (scatters) sound waves rather than absorbing them might be helpful, in some cases. That could include something like a bookcase, or an acoustic panel designed to act as a diffuser. The idea is simply to create an irregularity at early reflection points, so that sound waves won't ricochet from the walls in predictable, straight-line, patterns. As noted, for some listeners, a wider soundstage may make the sound seem to come from beyond the sides of the speakers.
Listeners need to make their own determinations of what they prefer in their specific rooms. People wanting to experiment with their preference for more ambient sound from side-walls, versus treating those early reflections, can try putting some sort of fabric on a wall temporarily, to test whether or not they like the result. Determining where to put a piece of test fabric, such as a blanket or a folded towel, or something with a surface irregularity, is fairly easy.
One person sits at the MLP, while another person holds a mirror on the side-wall. Where the seated person can see the speaker in the mirror is the first reflection point for that wall. And, that demonstrates where some type of temporary absorbing or diffusing material can go. If a mirror test results in subjectively better overall sound quality, for a particular listener, he now knows where to install nicer and more permanent acoustical treatments.
3. Center Channel:
The center channel (CC) is an extremely important component in an HT system, as it carries so much of the content in movies and TV programs. And, the placement of the center channel can be very important for dialogue clarity in both movies and TV. The front speakers are typically designed so that the tweeters are more-or-less at the ear level of a seated person. That is typically the case with both tower speakers, and with bookshelf speakers on stands. But, most center channels (except for those placed behind an acoustically transparent screen) are necessarily either below or above the display (or screen). And, they are typically horizontal speakers in order to facilitate those placement options.
[That horizontal placement necessarily means that most center channels, in mixed use rooms, will sit on some sort of cabinet. If possible, the speaker should not be enclosed within a cabinet, in order to avoid the same boundary interference that was discussed with the front speakers. The shelf a center channel sits on, however, is rarely mentioned as a potential issue with respect to SBIR.]
When CC's are placed above or below a screen, higher frequencies (which are highly directional) are not beaming straight at most listeners. As noted earlier, vertical dispersion, from both mid-range speakers and from tweeters, is usually restricted in order to avoid too many early reflections from floors. Consequently, when the speakers don't point more directly at listeners, that can interfere with overall sound clarity, and particularly with dialogue intelligibility.
(This is just speculation on my part, but I believe that where a center channel is not pointing directly toward us, we are still likely to be able to hear most of the lower frequencies in the human vocal range, and that would include most of the vowels. But, if the higher frequencies are attenuated, due to being so directional and pointing away from us, or if they are distorted by interference with a cabinet, or by early reflections, some sounds may be harder to hear. That would include some of the consonants, such as "B", "C", "D", "G", "P", "T", "V", and "Z", especially at the beginning or end of words.
Sounds from those consonants, and from other fricative sounds such as "F" and "S", which are produced at the front of the mouth, involve higher frequencies. Without them, the intelligibility of individual words can be lost. And, that may make it harder to understand what is really being said. That might especially be the case with softer or more rapid speech, where foreign accents are involved, or where background sounds are partially masking dialogue.)
It is advisable to make sure that CC's are pointed as much toward ear level as possible. Where a CC is above the screen, a shim placed under the back edge of the speaker will point it down more toward ear level. (Based on AVS Forum experience, however, it is harder to get clear sound from a center channel which is placed above a screen and pointing downward. And, it also a little harder to maintain the illusion that voices are actually coming from the characters on the screen.)
Where a CC is located below a screen, a shim placed under the front edge of the speaker will point it up toward ear level. Getting the center channel pointed more directly toward ear level can sometimes make a substantial difference with respect to audio clarity. And, when a center channel points directly toward our ears, our brains will usually adapt even more readily to the illusion that the speech we hear is coming from the mouth of the speaker on the screen.
Another placement issue that can affect both overall audio clarity and dialogue intelligibility, is the extent to which a CC is recessed on a shelf or within a cabinet. The potential for SBIR was mentioned just above, but there are also other issues with a CC placed inside a cabinet. As sound waves emerge in a cone shape, high-frequencies reflect off the top and bottom (or sides within a cabinet) of a shelf, causing what's called comb-filtering effects. That's a very jagged frequency response, with lots of little peaks and valleys, at higher frequencies, that interfere with sound clarity.
The result of those very early reflections from the shelf, arriving so close in time to the direct sound waves, can create a distorted sound and can especially interfere with our ability to hear dialogue clearly. Pulling the center channel forward so that the front of the speaker protrudes about an inch clear of the shelf or cabinet will help to prevent comb-filtering.
[Where it is possible to do so, it is always better acoustically to not enclose speakers (or subwoofers) inside a cabinet. It may be counterintuitive to think that enclosing a wood cabinet speaker, inside another wood cabinet, would negatively impact the sound. We may think that the cabinet would act to reinforce the sound and add more bass. It should reinforce the sound, from the standpoint of having sound waves bouncing around inside the display cabinet, or whatever the enclosure is. But, those sound waves which bounced off the sides, top, bottom, and back of that cabinet will reach our ears just enough later (or out-of-phase with the direct sound) to interfere with the clarity of the direct sound from the speakers. If possible, speakers should always be at the forward edge of a shelf, and if they can be taken completely out of cabinets, that is advisable.]
Irrespective of good center channel positioning, it can sometimes be a little difficult to hear dialogue in movies clearly. There are a number of factors that could contribute to a loss of clarity in general, or to dialogue intelligibility in particular. Sometimes, center channels are simply a little too weak to keep up with the other speakers in a system, or with the subwoofers. The CC may be the most important speaker in an HT system, as it is involved in playing at least about 80% of the content in a movie, and nearly all of the dialogue.
Heavy bass boosts can sometimes contribute to difficulty in hearing dialogue, especially where crossovers of 100Hz or higher are used for the CC. When a higher crossover is used, strong bass boosts may make voices sound deeper, thicker, and harder to understand. In that case, a better CC, which can utilize an 80Hz or 90Hz crossover, may be helpful. (A procedure known as cascading crossovers can also be helpful with dialogue clarity. It is explained in Section III-C.)
The use of DEQ (an Audyssey program) may also affect dialogue intelligibility, in some cases, as it slightly boosts the bass in all of the channels, including the center channel. And, it boosts the surround channels by about 1db for every -5db of master volume. The louder sounding surround channels could make it harder to hear the CC. (DEQ is discussed in detail in a later section.) Even without the action of DEQ, ambient noises, music, and special effects in a movie or TV soundtrack, may make sounds from the center channel harder to hear, and voices harder to understand. Many people prefer to boost the volume of their center channel a little, depending on the specific program, or on their specific circumstances.
4. Early and Late Reflections, and Room Treatments:
I have mentioned early reflections from side-walls that may involve the front speakers, and from cabinets and shelves that involve the center channel. There are other areas of the room that involve both the front speakers and the CC. If the floor between the front soundstage and the listening position is a hard surface, such as concrete, tile, or wood, there are almost certainly going to be some early reflections which can interfere somewhat with sound quality. In that case, an area rug in front of the listening position can be very helpful. Typically, people will use a foam rubber pad under the rug to prevent slipping, and to partially attenuate reflections from frequencies above about 1000Hz.
Coffee tables directly in front of a sofa can also be a source of early reflections, as high-frequencies bounce off the hard surface and reach our ears just behind the direct sound from the speakers. Any softening influences that can be added to the table can help. Even scattering some magazines on a table can help to disperse high-frequencies, so that they don't reflect directly toward a listening position. Remember that higher frequencies leave the speaker in a cluster, and they tend to travel in a straight line, so scattering (diffusing) them can also help to reduce the distortion we may hear.
Two other areas of a room can also be particularly problematical. First, as noted earlier, speakers radiate some sound waves backwards. If they are fairly close to a wall (perhaps within a foot or two), it can be advisable to put something such as an acoustic panel, or an oil painting, or decorative fabric behind the speaker. That will prevent early reflections from the wall from interfering with the clarity of the sound reaching our ears. If a dense acoustic (rockwool or fiberglass) panel is used, in a 1" or 2" thickness, it will typically only affect frequencies above about 240-300Hz. Foam rubber treatments, such as the egg crate versions shown in a video below, may not absorb frequencies below about 1000Hz, although they can also act as diffusers. But, enough of them in a room can have a significant effect, especially with high and mid-range frequencies.
If a listening position, such as a sofa, is within about 2' or 3' of a wall, it may be very helpful to put something on the wall behind the sofa. Ideally, something like an oil painting, or a tapestry, or an acoustic panel, would be used in order to disperse or absorb sound waves. Otherwise, those mostly high-frequency sound waves would bounce off the wall behind us, and into our ears, just a few milliseconds behind the direct-arriving sound from our speakers. Reflections from that back wall could also interfere with our audio clarity. Aside from hard surfaces, such as a floor or table, directly in front of a listening position, this might be the most important early reflection point when a listening position is near a wall.
This idea of reflections from front and back walls, and from floors and table tops, is probably worth expanding on a little bit. If there are reflections from a cabinet top holding a center channel, or from the floor, or from a table located within a few feet of a listening position, the result may be the same. That is because, in each case, the sound would arrive at a listener's ears only a few milliseconds behind the first arriving sound. (Again, sound travels at ~1ms per foot at sea level.)
The same thing would apply to reflections from a wall behind a speaker, or from a wall just behind the listener. In both of those cases, the reflected sounds would arrive just a few milliseconds behind the first arriving sounds. For instance, let's assume that the main speakers are pulled out 2' from the front wall. Some mid and high-frequency sound waves would travel backwards through the cabinet to reflect off the wall behind them. Their round trip of 4' to strike the wall and ricochet toward the listening position, would put their sound arriving ~4ms behind the sound waves that came from the drivers firing directly toward the listening position. That 4ms second difference could be just enough to create some audible distortion.
Distortion could also occur from sound waves reflecting from a wall behind a chair or a couch. If the couch were within a couple of feet of a wall, then sound waves would go past the couch, hitting the wall and ricocheting back toward the listening position. That could also create distortion, as the sound waves would still arrive at our ears well within that approximately 6ms window referred to earlier. (Again, when our brains can easily separate sounds, the first arriving sounds will seem louder, and we will tend to ignore the later arriving ones.)
I have observed that when someone is able to move the couch out about 3' or so from a wall, the problem is much less noticeable. In that case, the reflected sounds would be arriving a full 6ms or more behind the first arriving sounds. And, in some cases, that is just enough for the first arriving sounds to mask the later arriving ones. When in doubt, putting some acoustical material on the wall behind a couch is a good idea.
Examples of decorative products which can absorb higher frequencies, and which can be strategically placed in appropriate locations in a room, are all over the Internet. Here is an example of one such product:
AcousticART Panels with Custom Graphics and Images – Printed with YOUR OWN Photos or Art
* Reducing Reverberation (Ringing):
Most of what I have addressed so far would be classified as early reflections. But, as noted in the first part of this section, frequencies continue to bounce around a room until they simply run out of gas, and that can cause audible problems too. As noted earlier, in the discussion about distortion, we may particularly notice when higher frequencies do too much of that. And, the presence of many bare or hard surfaces in a room allows higher frequencies to continue to ricochet around the room for a much longer period of time. When a relatively bare room has a lot of excess sound energy in it, the result can sometimes be perceived as shrill or harsh sounding. That shrillness can often make it difficult to tolerate very high volume levels.
The same thing happens with low-frequencies in an untreated room, but the sonic effect of the longer time that it takes those frequencies to decay is very different. As noted earlier, low-frequency distortion is often described as boomy or muddy, where higher frequency distortion may sometimes be described as sounding too harsh or shrill. Both low and high-frequency distortion indicate a lack of clarity in the sound, although our reaction to them may be slightly different. I will start with a more complete discussion of higher frequencies as they are often the most noticeable and objectionable.
Listeners sometimes note becoming fatigued by their high-frequencies. Some speakers may be more likely to create that sensation than others, and some listeners may be more susceptible to that sensation than others. But, in either case, a very bare room will usually exacerbate the sensation of brightness or shrillness in an audio system. I have been in a listening room in a "high-end" audio store that sounded like fingernails on a chalkboard to me, at anything above a low volume level, simply due to all of the bare surfaces in the room. Some listeners, such as the employees at that audio store, can become so acclimated to a particular sound that they may not even notice the overall distortion they are hearing.
Some audio magazines may contribute to an impression that bare room surfaces are appropriate for "audiophiles" by publishing photographs of audio systems in very bare rooms. I have seen photographs of very high-end speakers, with equally expensive amplifiers, in rooms with all bare surfaces, including polished hardwood floors. The rooms were attractive and the speakers looked very nice in those photographs, but I would personally not have enjoyed listening to music in those rooms.
Whether or not to add any rugs, or drapes, or other softening influences to a room, is certainly a personal decision, and sometimes there may be important WAF (wife approval factor) considerations, as well. But, in general, bare rooms will promote distortion, which may be especially noticeable at mid and high-frequencies. It is important to note that ringing (meaning a longer decay rate of sounds within a room) is not something which will be visible from a graph of the frequency response of an audio system.
A waterfall graph (obtained through the use of REW) can illustrate where longer decay rates are occurring, but they can be both difficult to interpret and hard to correlate to what we actually hear. The physical reverberation time of sounds can be measured, as an R60 value, with the right equipment and technical understanding. But just as with a waterfall, we still have to correlate the reverberation time to our own listening preferences.
Where a listener lacks measuring equipment, but suspects that the bare surfaces in his room may be contributing to excessive high-frequency energy, a simple handclap test will help to determine how much of a factor the room itself is playing in the sound he hears. Standing at the listening position and simply clapping your hands sharply will tell you a lot about the room. If the sound of the handclap is prolonged, the reverberation we are hearing is called slap echo, or flutter echo, or simply ringing. Ringing interferes with our ability to hear individual sounds distinctly, because louder or first-arriving sounds linger for too long, drowning out harmonics of the initial sound, or the more subtle sounds which follow.
A very good example of this is an instrument such as a chime or a cymbal. There are subtle harmonics from the fundamental sounds produced by those instruments (and many others) which go way up in frequency. (A fundamental sound or frequency is a single note, but as stated earlier, all musical notes have harmonics which go up in frequency, at a somewhat reduced volume.) When the room itself exhibits ringing, we won't hear those harmonics as clearly, or at all. We will just hear the first loud fundamental sound, and the ringing in the room will drown-out the more subtle harmonics, which would otherwise cause the sound of the chime to linger in the air for just a moment, with the correct timbre, as the recording intended for it to.
This idea may seem a little counterintuitive at first, but what I am saying is that if the room itself has ringing (prolonged reverberation which would be almost like an echo in a larger space), then the natural timbre of an instrument, which is captured in a good audio recording, may be eclipsed by the inherent ringing of the room itself. And, the musical, or other sound, will be cut-off somewhat abruptly, and harmonics of that sound may be supplanted by room reverberations of the fundamental sound, which continue to reverberate in the room.
Other good examples of instruments to which this might somewhat subtly apply are the piano or other string instruments, or wind instruments such as the flute or clarinet. I suspect that most of us won't even be aware of what we might be missing unless we experiment--comparing very lively-sounding rooms, to less lively-sounding rooms, and listening for the differences.
The YouTube video I am linking below has a terrific illustration of the ringing phenomenon I have been describing. In this instance, the instrument is a snare drum, playing in a room with all hard surfaces, oblique angles, and no furniture. This is a deliberately extreme example, as most rooms will have a little more favorable geometry, some furniture, and some other softening influences.
So, in most cases, the ringing effect would be much more subtle than what we hear in the video. And, in some cases we might not really notice it without a before-and-after comparison. It would also typically take much less effort to reduce the ringing in a normal room, and we might not want to deaden the room quite as much as is illustrated in the video. (This is an issue that will be explored a little more below.) But, the video dramatically makes the point of what excessive reverberation in a room can do to distort mid and high-frequencies.
GIK Acoustics has a short article which builds on the example of the untreated room by illustrating differences at lower frequencies as well as higher frequencies. And, they show before-and-after waterfall graphs, along with a short audio track demonstrating the audible difference that adding full-range bass traps and other acoustical treatments can make:
Individuals who want to enhance the clarity of frequencies above the transition point may wish to investigate whether early and late reflections are a problem in their rooms. If so, simply adding some softening influences to the room may greatly enhance the quality and clarity of the sound in the room. Room treatments can be subtle and don't necessarily have to involve the use of acoustic panels. Area rugs, thick drapes, bookshelves filled with books, and other softening, absorbing, or diffusing influences can also be very effective with higher frequencies.
The YouTube video illustrated above above provides a great example of what excessive room reverberation sounds like with upper-bass and mid-range frequencies, using a snare drum. In my experience, any sound which is intended to linger a little is compromised by excessive reverberation. The more dominant sound lingers instead, and the subtle harmonics of the original sound are lost. High-frequency sounds, which have a certain degree of inherent brilliance, due to their harmonics, may become especially distorted when reverberation times are high. And, once heard and recognized for what it is, that type of distortion is very hard to unhear. I definitely notice it in rooms with relatively poor acoustics.
When ringing in a room has been addressed, a handclap test reveals a single fairly sharp sound, with relatively little lingering quality. And that, in turn, allows the full frequency range of individual instruments, and of individual sounds to be heard, without the room itself getting in the way. To what extent this is an issue in a particular room, for a particular listener, is another YMMV decision. As a general rule, it is probably better to add acoustical treatments and/or softening influences to a room gradually, listening as we go. Concentrating on the specific areas of the room, mentioned earlier in this section, would also be a good idea.
How much overall reverberation, or ringing, we prefer will undoubtedly vary from individual to individual. Most of us would probably want to shoot for a reverberation time (RT60) of somewhere between about 0.2 and 0.7 seconds, for our HT's. But, that is a very wide range in terms of perceived sound quality. (RT60 is an industry standard for the time it takes for a sound inside a room to decay by 60db.)
I have seen both 0.4 seconds (400ms) and 0.7 seconds (700ms) described as optimum HT targets from two different home theater designers. I believe that part of what makes the different recommendations confusing is both the goals and the preferences of the individual designers. A longer reverberation time (in the .5 to .7 range) would perhaps provide better music fidelity, but might also make movie dialogue slightly harder to distinguish. A shorter reverberation time (in the .2 to .4 range) might provide more dialogue clarity, but with a slight sacrifice in liveliness for music. Everything is a compromise!
But, that's where personal experimentation and personal preference come in. Finding the "optimal" compromise for a particular HT room is an extremely individualistic exercise, and that's why it can be helpful to go slowly and to listen carefully when adding room treatments. Another aspect of this is probably important to point-out. Room size and room geometry matter with respect to reverberation time. Larger rooms can support longer reverberation times than smaller rooms can, and that's one reason that the ideal reverberation time is shown as a fairly broad range.
Remember the illustration that showed the correlation between room size and perceived loudness. In that graphic, a 1500^3 room sounded +4dB louder at the same volume level (SPL) than a 5,000^3 room did. A small room in the 1,000-2,000^3 range would probably lend itself much better to a .2 second reverberation time than a 5,000-6,000^3 room would. The room geometry and the location of the listening area probably also make a difference.
The worst type of room, for reverberation, would probably be a small square room, such as a 12' by 12' by 8' room (1152^3). And, the reverberation issue would be intensified near the center of the room. (It could probably also be intensified, especially for bass frequencies, by lighter construction materials and a suspended wood floor.) It is likely that getting a reverberation time much closer to .2 seconds would be desirable in that room. A longer rectangular room (such as 12' by 20') or a room with an irregular geometry, could probably support a little longer reverberation time and still deliver comparable sound quality. And, a much larger room, such as mine at 6,000^3 with an irregular geometry, can support a reverberation time around .5 seconds or higher.
One final factor that could be worth mentioning is listening volume. Someone who never listens with a master volume above -20 or -15, could probably get by with less room treatment, and a higher R60 value, than someone who listens at master volume levels above -10. So that could also be worth considering as room treatments are added. In any event, it is clear that the decision as to how much room treatment to do, in order to achieve our personal audio goals, is an extremely personal decision.
The article which follows explains reverberation time in both simple, and more technical terms, as we follow the associated links.
Room & Building Acoustics
To get a sense of how we might determine the approximate reverberation time in our rooms, with a handclap test, the following series of short videos may be very helpful. In the first brief example, the RT is probably about 0.2 seconds (200ms). That room might feel very dead or dry acoustically--perhaps similar to a recording studio. The next one demonstrates an RT of 1.0 second. I would personally consider that room to be a little too lively for music listening and for HT. Somewhere between the first two examples (perhaps around 0.3 to 0.6 seconds) is where most people will probably want to be. Short of measuring our rooms with appropriately implemented acoustical analyzers, performing a handclap test should give us a good general sense of how we are doing when we add acoustical treatment to our rooms.
How to get a feeling for RT60 value
If we go slowly with our acoustical treatments, adding softening influences gradually, and paying attention to our own perceptions of sound quality, there should be no risk at all of over-treating a room in a way that compromises desirable ambient sounds or helpful reverberations. It is probable that different individuals will prefer slightly more, or less, room treatment, so two different people could probably address the same room in somewhat different ways. The main thing is to become aware of the ways that our rooms may be negatively affecting our sound quality, so that we may better please ourselves with the aesthetic/acoustic choices that we make.
A rule-of thumb which always made sense to me, for audio in general, was that a room which sounds comfortable for normal conversation is likely to sound comfortable for listening to recorded music, and perhaps also for watching movies. That is because, just as with normal conversation, the room itself wouldn't be adding or subtracting too much to the sound for music listening, or for movie watching.
A room which is so lacking in reverberation, that it isn't very comfortable for hearing normal conversation, at normal speaking levels, may also be adversely affecting the normal sounding timbre of voices, and especially of higher frequencies. (Remember the earlier musical terms that described "presence" and "brilliance" in frequencies above about 4,000Hz.) The same dullness that we might hear with normal conversation may be expected to carry over to what we would hear with recorded music.
Alternatively, a room which is so reverberant that voices sound abnormally loud or somewhat distorted for conversation, probably won't sound very good for music, or for movies, either. Musical instruments, playing higher frequencies, and vocals, may be particularly affected, as we will miss out on subtle harmonics. And, some sounds may become relatively shrill. Movie dialogue can also be especially affected by excessive room reverberation, as the center channel producing the dialogue is already competing with several other sound sources, even without excessive room reverberation to exacerbate the problem. So, in HT systems, excessive room reverberation may mask dialogue clarity in movies. In all cases, though, this is a YMMV issue, which individuals will need to resolve to their own satisfaction. Sometimes aesthetics, or even simple inertia, will trump acoustics, and that's fine too.
** Bass Traps:
For frequencies below about 250Hz or 300Hz, bass traps can be very helpful in reducing the overall decay rate in a room. And, as in the audio example that GIK used in an earlier link, reducing the bass decay rate in a room could make the bass sound clearer and less boomy. But, it takes thick and dense acoustic panels to affect frequencies below about 120Hz. Bass traps are acoustic panels which are specifically designed to absorb some longer bass wavelengths. As explained in more detail in Section VII, low-frequencies behave more sluggishly than higher frequencies do, pooling in areas where any two (or more) room surfaces meet.
For that reason, they are referred to as "standing waves". When standing waves collide at room boundaries, the low-frequency wavelengths either reinforce each other, causing peaks at certain frequencies, or they cancel each other, creating dips or deep nulls. Peaks at random frequencies can create a boomy, one-note-bass effect. Dips and deep nulls make it difficult, or impossible, to hear some bass frequencies.
As some of those excess standing waves are collected by the acoustic traps, random peaks and valleys in the frequency response are reduced, allowing other low-frequency sound waves to be heard more clearly. Boomy sounding bass is often the result of a loudness peak, at a particular frequency, that obscures other low-frequency sounds. On the other hand, bass traps can often allow bass to sound both clearer and louder than it did before, when significant cancellation was occurring. As demonstrated in that GIK audio track, bass boominess and an overall lack of clarity can also be attributable to the bass decay rate in a room. And, bass traps can help to remediate that.
Unlike other room treatments, used to reduce ringing at higher frequencies, it is probably more difficult to go too far in reducing the reverberation time of frequencies below about 200-300Hz. I believe that would especially be the case in a smaller room, involving relatively lighter construction and/or a suspended wood floor. Longer bass reverberation times in a small room could really obscure other frequencies, and sound extremely boomy to some people. So, I think that most people would probably be safe in adding as much bass trapping as they wanted to, although it is likely that not all of the bass traps would need to be full-range. Even there though, there could be some degree of personal preference involved in selecting the preferred amount of bass reverberation in a room. Going slowly, and adding treatments gradually, is still probably advisable.
Bass traps are typically at least 4" thick and are typically made of compressed fiberglass or rockwool. The ones with which I am most familiar have a plywood backing with a large hole cutout in the plywood. The hole works something like a Helmholtz resonator to attenuate low-frequencies. Then, the entire panel is covered with fabric. The best results are obtained when the thicker panels can stand-out away from the wall, with a gap between the panel and the wall behind them. I believe that using about a 4" gap behind a bass trap is usually recommended for very low-frequency absorption. The combination of the thicker acoustic panel, and the air space behind them can make bass traps somewhat effective down to about 50Hz or 60Hz, where the panel sitting flush against a wall, might only be effective down to about 120Hz.
Thick bass traps (4" to 6" thick) can be placed anywhere, but are frequently recommended for placement in corners, as that is nearly always a location where low-frequencies collect. The large corner traps are typically wedge-shaped to fit in corners, and may be about 12" deep at the point of the wedge. That 12" depth may consist of a 4" or 5" thick panel which faces out into the room, with two plywood side pieces, which form the triangular shape that fits into the corner. The entire corner bass trap is covered in fabric.
(Another type of bass trap doesn't involve a plywood panel. Instead, the entire 12" deep, wedge-shaped panel is composed of compressed rockwool. I understand that they can be as effective as the panel traps, which have air pockets behind them.)
Where a plywood panel is employed, the side pieces create a built-in air pocket behind the thick front-facing panel, so that the wedge-shaped bass trap can fit flush into the corner. As noted above, the thick bass traps, with air gaps behind them, can be at least somewhat effective down to about 60Hz. Bass traps are typically designed to also absorb mid and high-frequencies, in addition to the low-frequencies for which they were designed, unless otherwise specified. Those "broadband" bass traps can, therefore, also reduce ringing in mid-range and treble frequencies.
Sometimes, where bass traps can be partly effective with mid-bass frequencies (defined as about 50Hz to 120Hz), they can allow room correction to complete the job of smoothing-out peaks and valleys in the frequency response. (Room correction alone, is not always effective in doing that, nor can it meaningfully affect the overall decay rate in a room.) As with the higher frequencies, there is a good deal of personal preference involved in deciding how much bass trapping is required. Sometimes adding just a few bass traps enables a listener to achieve a clearer bass sound, without the sluggishness or boominess that he heard before.
There are a number of good sources for these more specialized acoustic panels, and they can also be DIYed. Most of the acoustic panel makers will offer free room treatment advice. I am showing a link to one well-known maker, but Ethan Winer's RealTraps, and ATS Acoustics are also good sources.
GIK Acoustics - Bass Traps with FlexRange Technology©
5. Location of the MLP:
With respect to the question about where to locate a main listening position (MLP), there are two factors to consider. One factor involves room modes, which only affect bass frequencies, and which mainly affect frequencies below about 250-300Hz. We will notice those room modes most with our subwoofers.
Typically, being at the exact center of a square or rectangular room is the worst possible listening position. 1/4 or 3/4 room length can often work well, but 3/8 or 5/8 of room length can sometimes be a little better. Those 1/8 of room length positions are theoretically "ideal" MLP positions, although measurable differences between those and 1/4 length positions can sometimes be minor.
If REW (an independent measuring program) or something such as the Audyssey App is used, it may be possible to differentiate between an MLP which has peaks at some frequencies, and one which has nulls (cancellation) at some frequencies. If it possible to differentiate between peaks and nulls, then a location with peaks will be easier to deal with than a listening position with nulls. For instance, an automated system of room EQ, such as Audyssey, can reduce peaks in frequency response. But, it can't do much with respect to deep nulls.
Again, experimentation is the key to discovering what works best, and as noted above, REW can be your friend in that process. Speaking generally, larger rooms tend to be a little more forgiving about room modes, and consequently about listening positions and subwoofer placement, than small rooms. A large room, in this context, might be about 4,000^3 or greater.
The second factor in determining a preferred listening position has nothing to do with bass frequencies. Instead, it involves early reflections of mid-range and high-frequencies. For example, in a large room, a listening position might be at 7/8 of room length and still be far enough away from the wall behind the MLP to avoid early reflections from that wall.
Ideally, you would want to be at least 1 1/2' to 2' from the rear wall, and 3' might be better. Otherwise, frequencies from about 1,000Hz and up may reflect from that wall in a way that garbles (distorts) the sound. That is because our brains can't easily separate sounds of equal volume, that arrive too close together. This was discussed in some detail in the previous subsection.
In a small room, however, it can be difficult to have a listening position which is 5/8 or 3/4 of room length (for bass frequencies) and still be 2-3' from the rear wall of the room, in order to prevent early reflections of higher frequencies. In that case, putting an absorbent panel on the wall behind the listening position can resolve the problem of early reflections.
Although most people on the subwoofer forum probably concentrate more on bass frequencies, it is important to understand that there can be two different objectives involved with respect to the location of the MLP. Sound quality for the higher frequencies is also an important objective.
Section I-C: Room EQ and Calibration Techniques
Systems of automated room EQ, such as Audyssey, measure the frequency response within a listening area, and then set filters (control points) at specific frequencies to equalize sound pressure level across the entire frequency range. Automated room EQ is generally believed to be helpful with bass frequencies (<500Hz) and especially helpful below the transition frequency in a room. The extent to which systems of room correction can correct higher frequency distortion in a room, caused by setup issues or by fundamental room issues, is a more controversial question.
In my opinion, this question can only be answered by individual listeners on a case-by-case basis. If the overall audio quality sounds better with Audyssey on, then the room correction is successful in that specific instance. And, if the overall sound quality sounds better without Audyssey, or some other form of room correction, then I see that as strictly a user preference issue. (Before writing-off a system of room EQ, such as Audyssey though, it is a good idea to experiment extensively with settings, such as DEQ. A number of user-controlled settings can influence the potential sound quality for a particular listener in a particular room. They are discussed throughout the Guide.)
* Note: Some systems of automated room EQ allow users to limit the use of room correction above a certain frequency, such as 500Hz. That is strictly a user preference issue. The Audyssey app, which is available with newer model Denon/Marantz AVR's/AVP's, has that feature, along with other user adjustability options. There is a thread devoted to that app, so I am not going to address it in the Guide. Instead, I am going to explain ways to maximize the sound quality of Audyssey calibrations in general. Readers who want to learn more about the use of the Audyssey app are encouraged to consult this thread:
MultEQ Editor: New App for Denon & Marantz AV...
There are some things that we can do initially, to enhance our chances of improving our audio quality, when we do a room correction calibration with a system such as Audyssey. First, we have to understand that there may be a limit to what Audyssey can do to correct pre-existing problems involving improper speaker setup, or too many early and late reflections. Better speaker setup, and even a modest use of room treatments, may augment what room correction can do to improve the sound quality.
Second, we need to understand that, in some cases, room correction can actually create distortion (or exacerbate it, if it already exists) if we don't observe good calibration technique when we EQ our rooms. So, this section of the Guide will offer some general tips which may assist users in getting better calibrations. Although the general focus will be on Audyssey, some of the tips may apply to other systems of automated EQ as well.
We need to realize that measurement microphones, such as the one employed by Audyssey, do not "hear" sounds in the same way that we do. To start with, the Audyssey measurement microphone is far more sensitive than our own hearing is. Small variations in volume at certain frequencies, which might completely escape our notice, will be picked up by the Audyssey microphone. And, Audyssey will try to "fix" them, even if they don't really need fixing, and, even if we would already have been able to hear those frequencies with subjectively good sound quality.
For instance, if speakers are toed-away from the MLP a little too much, Audyssey may detect that the high-frequencies are a little too low, in comparison to the frequencies which are not quite so directional, and Audyssey may boost those high-frequencies accordingly, causing some shrillness. The reverse can also be true, when some types of tweeters are pointed too directly at the MLP. For example, some manufacturers (or experienced product users) may recommend that specific front speakers not point directly at the MLP. That can especially be the case with some horn speakers, which have high sensitivity, and which may sound bright compared to other speakers.
In either case, Audyssey may exacerbate a barely noticeable (or completely unnoticeable) but pre-existing situation, by trying to do too much correction in a frequency range where a lot of correction may not be required for an individual to hear subjectively good sound quality. I think that Audyssey can actually be used as a kind of test instrument to help users discover how to point their speakers in just the right direction to maximize sound quality, although the resulting differences in SQ may be subtle.
I found that to be the case in my own room. A little less toe-in gave me more harshness in the upper frequencies, post-Audyssey. A little more toe-in, and Audyssey did less to affect those higher frequencies, resulting in better, more natural, sound quality. In my particular room, and with the specific setup and adjustments I use, Audyssey simply sounds better on that off for the full frequency range, although the most noticeable difference is still below 500Hz.
What I am saying here is that if we wish to fully benefit from automated room correction, and if we are really serious about the sound quality in our rooms, then it may be necessary to experiment a bit with speaker positioning, with appropriate softening influences, and then with subsequent Audyssey calibrations, in order to achieve the best overall sound quality that we can. It is a little bit of extra work initially, but the long-term results can be worthwhile.
We shouldn't just assume that the Audyssey microphone will hear exactly what we hear, because it won't. The sensitive omnidirectional Audyssey microphone will "hear" things that we won't. And, our brains will filter and influence what our ears do hear. We also shouldn't assume that it will EQ our systems in accordance with our personal listening preferences, if we don't experiment a bit with our speaker placements and our EQ technique when we are running our Audyssey calibrations. It's really just trial-and-error to find out what works best for a particular listener in a particular room.
[I should note here that my intent in writing this is not to make people obsessive about their speaker placements, or about any other factor being discussed. The real intent is just to inform people that factors such as speaker toe-in can potentially impact an Audyssey calibration. If the initial Audyssey calibration sounds good, then don't worry about it. If, over time, someone wonders whether any further improvements can be made, typically to the higher frequencies, then some additional experimentation might be helpful.]
A second example of the difference between what we hear, and what Audyssey hears, concerns the nature of the omnidirectional Audyssey microphone. The Audyssey microphone hears sounds equally in all directions, but we don't. The pinnae (flaps) in our ears funnel sounds into our ears from the front and from the sides. But, they partly block and deflect sounds coming from behind us. So, early reflections from a wall behind us are going to be noticed far more by the Audyssey microphone, than they are by our ears. And, in trying to over-correct those early reflections, Audyssey may actually contribute to the distortion we hear.
[As a general rule, we don't want Audyssey to become extremely busy in the high-frequency range. Reflections from hard surfaces, bouncing into the Audyssey microphone at close range, can cause Audyssey to set too many control points at high-frequencies, causing additional distortion which may sometimes be characterized as comb filtering. The more that we understand why Audyssey might be doing things to "correct" mid-range or high-frequencies, the more that we can enjoy the overall benefits of room correction for low-frequencies, without adversely affecting our higher-frequency sound quality.]
The nature of the omnidirectional microphone is why Audyssey advises keeping measurement microphones at least 18" away from a wall or other hard surface. Perhaps an even better example of the difference between the way we hear, and the way the microphone hears, involves chair or sofa backs. Most chairs or sofas in our HT's or mixed-use rooms have relatively smooth surfaces. Some of them have fairly firm leather surfaces.
Those smooth or firm surfaces reflect high-frequency sound waves directly into the Audyssey microphone, in a way that they never could if we were actually sitting there. And, the sounds from the back of a sofa would be sufficiently attenuated by our pinnae (ear flaps), and would reach our ears so simultaneously with the direct sound, that we would never hear them. But, the omnidirectional Audyssey microphone would hear them, and in trying to correct something that didn't need correcting, it could introduce comb-filtering (high-frequency distortion) into the sound.
One way to avoid that problem would be to keep the Audyssey microphone at least 18" away from a chair back. But then, we wouldn't be measuring where our ears are, and that could negatively impact our calibration. A better solution is simply to drape a fluffy blanket over our chair backs during calibration. That will enable us to get our microphone within about 4" or 5" of the chair back, and where our ears actually are as we listen. At the same time, that will prevent high-frequency sound waves from bouncing into the mic from very close range. And, Audyssey will leave those spurious high-frequency reflections alone, concentrating its EQ resources on broader areas of the frequency range. (Chris Kyriakakis, the creator of Audyssey, has endorsed this solution.)
Audyssey employs a system of fuzzy-logic weighting to average the results from either six or eight microphone positions (depending on the Audyssey version). In general, I believe that the more we can give Audyssey more consistent measurement results to work with, the more that we can achieve a smoother frequency response, and consequently improved sound quality. This is something that Chris Kyriakakis commented on in response to a question. He suggested that the more uniform the sound is within a measurement area, the more uniform the Audyssey EQ is likely to be. And, the more uniform the Audyssey EQ is, the more likely it will be to provide good sound quality to a larger listening area. (We should recognize that no system of automated room EQ is likely to be able to EQ an entire room. The smaller the listening area we are trying to EQ, the better our resulting sound quality is likely to be.)
** The examples above illustrate one aspect of microphone placement, within a more consistent measurement area, but general microphone placement is also a factor. We typically want to have our microphones at ear level, even if not all of our speakers are at that same height. Keeping the mic at roughly ear level seems to be consistently important. Some users, including myself, have achieved good results by taking just a couple of measurements 2" or 3" above ear level. We typically do not want to go behind a chair back with any of our measurement positions, unless we are deliberately trying to EQ for a second row of listening chairs. And, even then, it would be a good idea to experiment both ways--going behind the MLP, and not going behind it. It helps to recognize that Audyssey is simply trying to EQ a fairly uniform listening area, and not individual seats. Other than mic position 1, which is typically centered on the primary listening chair, the mic positions don't need to correspond at all to any actual seats.
Readers interested in why it might not typically be a good idea to measure behind the main listening position, if no one is actually sitting there (and often, even if someone is) may wish to read a post from later in the Guide thread. It adds a little more detail about the issue:
Guide to Subwoofer Calibration and Bass Preferences
It is a very good idea to use a boom microphone stand with an extendable arm for Audyssey calibrations. That allows the base of the stand to remain on the floor, while the swing-out arm allows the microphone to be positioned exactly where the listener wants it. If something such as a camera tripod is used, mic placement can be much more difficult, and the heavy mass of the tripod can interfere with the calibration due to secondary reflections from the tripod. If a tripod is placed in a chair or on a sofa, to facilitate mic placement, vibrations from the furniture can be passed up through the Audyssey microphone. Whether that would always make a significant difference in the calibration is somewhat debatable, although there have been some good before-and-after examples where it definitely did make a difference.
Another issue with camera tripods is the bulk of the tripod itself, positioned directly below the microphone. There can also be spurious reflections from the base of the tripod, which can interfere with the accuracy of the calibration. For both Audyssey and for REW, boom mic stands work much better. And, for the small cost involved, I think that it makes sense to use a much better and more stable stand than the cardboard one that Audyssey provides.
The type of stand I am recommending provides for much more exact mic placement, and better repeatability for calibrations. That repeatability is important, once someone has found a mic pattern that works well for his room and equipment. There are a number of different stands that can work, although I have heard that some of them which come with adapters included, are flimsy and don't stand-up well. The one that I am linking below is sturdy and durable, but it does require a separate adapter to hold the Audyssey microphone. I am also linking two different adapters. Either adapter can work.
Amazon.com: On-Stage MS7701B Tripod Microphone Boom Stand: Musical Instruments
Amazon.com: On Stage CM01 Video Camera/Digital Recorder Adapter: Musical Instruments
Amazon.com: On-Stage MY200 Plastic Clothespin-Style Microphone Clip: Musical Instruments
*** As a general rule, it is a good idea to measure smaller areas, as opposed to larger areas, for the reasons cited above. We want our measurement area to be large enough to accurately represent the binaural nature of our hearing. (For bass frequencies which are low enough to be non-directional, hearing them with just one ear may be sufficient.) For all frequencies though, we at least want to measure in a fairly large circle forward of and out to the sides of our heads. Fairly large, in this case, could be a circle about 18" to 24" in diameter.
But, we may not want to measure such a large area that we present Audyssey's fuzzy-logic weighting system with too much anomalous information. The more uniform the frequency response is, within a measurement area, the better the resulting calibration is likely to be. Patterns that vary in size from as small as about 6" to 12" out from the MLP (mic position 1), to as large as about 24" to the side and forward are typically used. I would not generally recommend going further out to the sides, or forward more than about 24" from mic position 1.
It is interesting to note that, in the last couple of years, Audyssey has revised it's owner instruction manuals to recommend a smaller microphone pattern than they used to recommend. They used to recommend 3' to 4' out from the MLP. I believe that they now recommend about 2' or less. Their revised recommendations seem to parallel the experience of many Audyssey users, who discovered that smaller microphone patterns often resulted in better sound quality, over a wider area, than large mic patterns did. That is consistent with my own experience, and with that of a number of others on the Audyssey thread.
But, I suspect that finding an optimum microphone pattern is at least somewhat room and system dependent, so I suggest that interested users experiment in an effort to discover the specific microphone pattern which produces the best sound quality in their rooms. Once they find a mic pattern that they really like, I recommend that they write it down, or draw it, so that they can return to it for future calibrations. Sometimes, fairly subtle differences in microphone placement can yield significant differences in the resulting sound quality.
**** For people who are looking for some preliminary guidance in selecting microphone positions, the following visual aid is offered. This roughly 2' by 2' pattern is one that a number of people have successfully used. But, it is only shown as a starting point and not as a specific recommendation. People still need to experiment to discover what pattern works best in their particular circumstances.
In this pattern, mic position 1 is about 4" or 5" in front of a blanket covered chair (or couch), the center of which is the MLP. Remember that MLP stands for main listening position, and mic position 1 is always, by definition, the MLP. The MLP can be the center of a couch, or the center of a chair, depending on the specific room. For purposes of the illustrated diagram, mic position 1 is right between your eyes (and ears) and about 4" or 5" way from the blanket covered surface of a chair or couch. The mic is at a height which approximately corresponds to the center of your ear canal, as you would sit when listening to music or watching movie. That is what we mean by ear height.
It may be important to to note here that mic position 1 is used to set volume levels and timing (distance) for all of the channels. In order to accomplish that, Audyssey only uses a portion of the full bandwidth sweeps in mic position 1. It uses the 30Hz to 70Hz bandwidth to set levels for the subwoofers in mic position 1, and it uses the 500Hz to 2,000Hz bandwidth for the other channels. Full bandwidth sweeps of 10Hz to 22KHz are employed for all of the mic positions in order to set EQ filters. It is important to keep all of the mic positions fairly close together in order to insure that Audyssey's system of fuzzy logic weighting is presented with somewhat uniform measurement information.
Positions 2 and 3 are out to each side of 1 by about 10" to 12". Positions 4 and 5 are straight out in front of 2 and 3 by about 20" or 24". Number 6 is in a straight line out from 1, but this time only about 14" to 18" away. All six of those mic positions are right at ear height. Positions 7 and 8 are in fairly close to the chair back--perhaps only about 6" away from the blanket and about 6" out to the side of mic 1. (That clusters some mic positions very near the head, and where the ears on each side of our heads are located.) Both of the last two positions can be raised up by 2" or 3" above ear level. In this particular mic arrangement, none of the mic positions go behind the chair.
2--------1---------3
-----7--------8-----
---------6-----------
4-------------------5
The specific order of the mic positions is not important, so after mic position 1 (which is always the MLP) the numbers assigned are arbitrary. Users can follow the diagrammed positions in whatever numbering sequence works best for them. It is only important to keep the mic level (so that it points upward) and close to ear height for at least about the first six positions. People who have a version of Audyssey which only uses six mic positions might wish to eliminate 7 and 8 from the diagram shown, or they could experiment with an even more compact configuration for their six. Experimentation is the key to finding a result which pleases the individual user.
Section II: Audio System Calibration and Subwoofer Levels
This section explains how Audyssey calibrates and EQ's our audio systems, and offers some advice for the best methods to boost our subwoofers. It also explains what Dolby/THX Reference is and how that standard specifically relates to the subwoofer boosts we may prefer.
The subsections in Section II are as follows:
II-A: Audyssey Calibration and Dolby Reference
II-B: Why We Add Bass After Calibrations
II-C: Where and How to Add Bass
II-D: Master Volume Levels and Sub Boosts
II-E: Gain Settings and Maximum Output
Section II-A: Audyssey Calibration And Dolby Reference
Audyssey Calibration:
Audyssey is a room correction software program designed to reduce room/speaker interactions which may adversely affect audio quality. Once we put a speaker or a subwoofer inside a given room, it's native frequency response changes, depending on a number of factors, including its specific placement in the room. Audyssey attempts to remove undesirable room influences by setting control points to even-out the frequency response, so that we don't have large dips or peaks in sound pressure level (SPL) at certain frequencies. Audyssey's goal is to make all frequencies approximately +/- 3dB from a standardized calibration SPL of 75dB.
There are multiple versions of Audyssey which have been introduced over the years. The newest and best version of Audyssey room correction (not counting the Audyssey App which simply permits more user control of what is being EQed) is Audyssey XT-32. The following table shows the various versions of Audyssey.
The versions are shown relative to 'X' filters in 2EQ, where 'X' represents 8 filters in that very early version. (We frequently speak of the numbers of filters used in room EQ. But in reality, there is only one filter per channel, and each filter has 'X' number of control points that it can employ.) The latest version of Audyssey, XT-32, has 4,096 control points available per channel. That is 8 x 512.
Audyssey, in all current versions performs its room correction by sending 75dB test tones to each of the channels in an audio system, and by then measuring the frequency response for each channel, from 10Hz to 22KHz. Once the measurements have been completed, Audyssey calibrates the results, using a system of fuzzy-weighted logic. It then sets control points at individual frequencies, or groups of frequencies, in order to correct for peaks and valleys in the sound. Once the control points have been implemented, the room EQ is complete.
* When Audyssey, and other systems of automated calibration, perform a system calibration they ignore all prior settings, such as the master volume level, trim levels, distances, and crossovers. Once the calibration is complete, the master volume level will typically return automatically to the volume that the AVR was on prior to the calibration.
It is important to note that all measurement microphones, including the Audyssey microphones, have an inherent error factor of anywhere from +/- 1.5dB to +/ -3dB. (The Audyssey microphones have an error factor of +/- 3dB.) That includes very expensive calibrated microphones used to measure SPL or frequency response, such as the UMIK-1.
If you measure the Audyssey test tones, or your post-Audyssey SPL, with a calibrated microphone, you may find that your SPL measures two to three decibels above or below 75dB. I think it would be fairly rare for two different microphones to measure exactly the same, in any case. The important thing is that all of the channels in your system are playing the same SPL, as measured at the MLP, and Audyssey will be quite accurate in that respect.
(It is not a good idea to double-check the post-Audyssey volume settings for the various channels, using your AVR's internal test tones. Instead, it's much more accurate to use external test tones, through a test disc. The internal test tones in Denon/Marantz AVR's bypass the filters that Audyssey set for the various channels. The test tones are in a completely different software program than the EQ program. It would not be uncommon for there to be a difference of a decibel or two between a calibrated speaker level and the uncalibrated one found in the test tones. The test tones are intended only for manual adjustments in trim levels.)
It is important to understand that two separate actions are performed during a calibration process. First, the system will be calibrated to Dolby Reference using 75dB test tones. All channels (including subwoofers) are set to have the same sound pressure levels, as measured at the main listening position (MLP), and all sounds are set to arrive at the same time, via the distance settings. Preliminary crossovers will also be set during this process. (As noted in the Cliff Notes, the preliminary crossover set by an AVR is not a recommendation. It is actually just an observation, as the AVR uses it's own default programming to make the preliminary crossover settings.)
The second thing that will occur, during the calibration, is the actual room EQ process. In that process, Audyssey will measure the frequency response from various microphone positions, and will use a system of fuzzy logic weighting to set filters for all of the channels, in order to remove some of the peaks and valleys in the frequency response that inevitably occur whenever transducers (speakers or subwoofers) are played inside a typical home theater or mixed-use room.
It is important to distinguish between the two processes. The initial calibration process insures that equal volume levels, from all of the channels, will arrive at the MLP at the same time, and it calibrates each of the channel volumes to a Dolby/THX Reference standard. The second part of the calibration process is the room EQ process, which sets filters for all of the channels, in an effort to improve the overall sound quality in the room.
The room EQ software program, and the filters (control points) it applies to an audio system are independent of the various AVR settings. So, once the calibration is complete, changing any settings will not change the room EQ that Audyssey applied to an audio system. This is a recurring question, so I have emphasized it here. AVR setting changes do not affect the room EQ that Audyssey applies. And, Audyssey can be turned off, and then turned back on again, as often as we like. Once Audyssey is turned back on, the same room EQ will be applied. (Using the Pure Audio mode will disable Audyssey. When Audyssey is disabled, features such as DEQ and Dynamic Volume are not available.) If significant changes are made to the room, however, or to any of the speakers or subwoofers (new locations, for instance) a new calibration should be performed.
One of Audyssey's goals, in any Audyssey version, is to set the volume levels of all channels in a system, including subs, to 75dB, as measured at the MLP, by the calibrated Audyssey microphone. The MLP is microphone position 1, by definition, wherever the user chooses to place the microphone. And, that point in space is where Audyssey will set timing (distance) and trim levels (volume levels), for all of the channels, to coincide.
The .1 in a 2.1, or 5.1 (or larger) audio system is the LFE (low frequency effects) channel, and that .1 designation has nothing to do with the number of subwoofers in a system. The .1 designation was originally selected because the LFE channel only plays a fraction of the total frequency range of an audio system. (The LFE channel is explained in a little more detail in Section III.) Where there are subwoofers configured in an audio system, Audyssey will measure and calibrate all of the subs in a multi-sub system together, so that their combined SPL is 75dB. Whether there is one subwoofer, or there are many subwoofers in an audio system, the combined sound of the subwoofers as a whole will be set to 75dB.
[The difference between subwoofers and the .1 LFE channel may be a little confusing at times to all of us, so I decided to add some clarification to this section. Most people reading the Guide will have subwoofers connected to sub outs in their AVR's or AVP's. Those AVR's and AVP's are designed to play Dolby 5.1 programming, which contains the .1 LFE channel, briefly described above. But, it is possible to connect subwoofers to some stereo amplifiers. We can also listen to 2-channel content, or watch older movies (made prior to the development of Dolby 5.1, which occurred in 1992) on our HT systems. However, the audio system is only playing LFE (.1) content, which is recorded 10dB louder than the regular channels, when 5.1 or higher material is being played on an AVR or AVP.
For any other program material, the subwoofers only play content contained in the regular channels (such as 2-channel content) even if the content is upmixed using a surround mode such as Dolby Pro-Logic (PLII). When subwoofers are employed for non-native 5.1 material, they simply enhance the bass in the regular channels by playing frequencies below the crossovers assigned to those speakers. It is important to understand the distinction I am making, because all of us tend to equate subwoofers with the .1 LFE channel, and they are two different things.
The .1 channel is for low-frequency effects content (special bass effects), recorded 10dB louder than other content in a 5.1 soundtrack. Subwoofers are transducers specifically designed to play bass frequencies, below crossover points, in the regular channels. And, in addition to that function, they also play .1 LFE content, whenever a 5.1 program is played through an AVR or AVP. It is worth mentioning that the .1 in 5.1 or 7.1 doesn't refer to the number of subwoofers. The .1 designation was used because the LFE channel was assumed to be playing about 1/10 of the total audio content in a movie.]
The trim levels and distances (timing) for all of the channels (including the subwoofers) will be set at mic position 1, from the initial test tones at that first mic position. Audyssey will disregard any previous settings and set levels, distances, and crossovers from scratch, whenever a new calibration is run. As noted, trim levels and distances for all of the channels are set based on microphone position 1. Crossovers are set after all of the test tones are completed, based on a fuzzy-weighted average of all of the microphone positions in a calibration.
The sweeps which Audyssey uses during its calibration process cover a frequency range from 10Hz to 22KHz. (A recent Audyssey zendesk answer stated that the sweeps go up to 24KHz. It may be that Audyssey's sweep range has changed in recent years, but if so, it won't make any difference whatsoever for the frequencies which humans can actually hear. So, I will keep using the range of 10Hz to 22KHz for the Guide.)
All trim levels and distances are set before Audyssey adds control points to the channels. And, all internal test tones, which govern those trim levels, remain independent of the EQ which Audyssey performs. The Audyssey EQ process occurs automatically once we tell Audyssey to "Calibrate". Audyssey takes the data from all of the mic positions and applies its fuzzy logic weighting, as previously described.
[The sweeps that Audyssey uses, during that calibration process, provide an interesting illustration of the way that our hearing works. The sweeps (broadband test tones) used for the regular channels sound much louder than the sweeps used for the subwoofers. But, all of the sweeps are playing at the same 75dB volume.
All of the sweeps, for all of the channels, are playing the same frequencies, but we aren't hearing them the same way. At a volume level of 75dB, the subwoofers can't go very far up into our normal hearing range with audible sounds, and that will make them seem much softer compared to our speakers which are playing sounds in our normal hearing range.
We may also hear a difference in the test tones, with larger speakers, which may sound lower in tone than smaller speakers do. That would be a product of the low-frequency capabilities of the speakers. And, we may also hear some difference in loudness between more distant speakers and closer speakers. Inside a room, bass frequencies lose about -3dB of volume for every doubling of distance. Frequencies above about 300Hz lose -6dB for every doubling of distance, due to the fact that the higher frequencies are not benefiting from boundary gain (from the walls, floor, and ceiling) in the way that bass frequencies are.
But, the biggest difference in both tone and loudness will be between the sweeps for the regular channels, and the sweeps for the subwoofers. The difference in the relative loudness between the sweeps for the regular channels, and the sweeps for the subwoofers, is the difference between the way we hear frequencies in our normal hearing range, and the way that we hear low-frequencies. Frequencies in our normal hearing range of about 500Hz to 5,000Hz sound much louder than the frequencies played by our subwoofers.
The test tone sweeps give us a good example for why most of us need to add more bass, once all of our channels are level-matched. That is especially true as our overall listening levels drop, because low-frequencies drop-away faster, relative to those in our normal hearing range.]
As noted above, the first microphone position is also used to set distances for all of the channels. Audyssey sets distances for the channels based on the time it takes for the sound to arrive at mic position 1. Subwoofers have their own internal amplifiers, and their own internal processing, which typically delays the timing of the sound arriving at the MLP. Audyssey compensates for that delay by setting subwoofer distances as longer than the physical distance from mic position 1. Setting a greater distance for a channel causes the AVR's internal programming to speed-up the arrival time of the sound (and vice-versa). In the subwoofer's case, speeding up the signal, with respect to the sound from the other channels, allows the sounds from all of the channels to arrive simultaneously at the the MLP.
Dolby/THX Reference:
When Audyssey finishes, all channels in the system (including all of the subwoofers' combined SPL) will play at the same volume, at the MLP, as determined by the calibrated Audyssey microphone. And, when all channels in a system are playing at the same volume, the sound at the MLP will be approximately in balance with what the film mixer intended whenever a movie is played at "Reference" volume, which is 0.0 master volume. However, when the listening level is lower than about -5 MV, most people will not hear bass frequencies quite as well as other frequencies, or quite the way that a film mixer intended for them to be heard in equilibrium with the other frequencies in the film.
The Dolby, or THX Reference standard is intended to provide some degree of uniformity in the maximum volume levels of movies, and is intended to provide a way for commercial cinemas and home theaters to make sure that their audio systems correspond to what film mixers intended for them to hear. The Reference level is capped at a peak volume of 105dB for the regular channels, and 115dB for the LFE (low frequency effects) channel. Those were considered maximum safe listening levels for movies. Most of those peak volume levels of 105dB for the regular channels and 115dB for the LFE channel would be for very short durations.
The LFE channel enables additional low-frequency effects, below about 120Hz, to be mixed into a 5.1 movie (or music) track. A system that is calibrated to Reference will, if it has the capability to do so, automatically play those 105dB and 115dB peaks when the master volume control is set at 0.0. It is important to note that not every system is capable of playing those 105/115dB levels, nor would those volume levels be found in every movie. (The .1 LFE channel is explained in some detail in Section III-B.)
As stated, Dolby/THX Reference is just a guideline that provides a degree of standardization for the way that movies are recorded, and for the way that commercial cinemas and home theaters are calibrated. But, all the Reference standard really does is to establish maximum volume levels for the regular channels (105dB) and for the LFE channel (115dB). There is no specific uniformity with respect to the average volume level of movies, and there is no specific uniformity with respect to how frequent, or how sustained, crescendos of up to 105/115dB will be in a particular movie. This is an important point to understand!
Some movies may be much louder than other movies, both in terms of the average volume level of the film, and in terms of whether the film actually hits crescendos at the upper limits of the Reference standard. Some movies may never hit the upper limits of the standards, and some movies may hit those upper limits again and again. That is entirely at the discretion of the director and the film mixer. That is easy to understand if we compare blockbusters to light romantic comedies. But, even among different blockbusters, or among action films in general, there may be volume differences. And, there may be profound differences in how low in frequency the movies goes. So, overall volume, peak volume, and very-low-bass volume, can all be variables among different movies.
Once an audio system has been calibrated to Reference, how much or how little of the total SPL capability of an audio system is actually employed, is entirely a matter of personal preference. Some people prefer to listen at much louder levels than others do. It is important to understand that Audyssey will set all of the channels in a system to play at equal volumes at the MLP. And, when the master volume is at 0.0, the audio system will be playing Reference volumes, if it is capable of reaching those sound pressure levels. But, it is up to the individual user to decide how loudly he wants his audio system to play, via the master volume control.
That same principle also applies to our subwoofers. Sometimes, people will add a second subwoofer (or a larger subwoofer) to a system, and then question why the bass doesn't sound any louder after the system is recalibrated with the greater amount of subwoofer headroom included. Having multiple subwoofers increases our potential bass headroom, so that we can hit higher bass SPL's. But as noted earlier, the combined SPL of our subs will always still be calibrated to 75dB, so that the combined volume of the subwoofers will be equal to the volume levels of the other channels in our system. In order to actually use any additional bass headroom we may have available, we either have to play our audio system at higher listening levels than we were before, or we have to increase the volume level of our subwoofer(s). Or, we can do both.
At one time, test tones were employed at the "nominal" average Reference volume of 85dB. (I use the term nominal average volume, since there really isn't an actual average volume level for movies. Film mixers just record movies at the volume level that seems appropriate to them, although some TV channels do impose strict standards on average and peak volume levels.) The 85dB number was selected when the Dolby/THX Reference standards were developed, to represent a hypothetical average, with 20dB of headroom above that for the regular channels, and 30dB above that for the LFE channel. That same 85dB number was chosen as a uniform calibration number. And audio systems were calibrated to Reference at 0.0 master volume with 85dB test tones.
However, the original 85db test tones which were used were uncomfortably loud for most people, so most systems of automated or manual calibration, including Audyssey, converted to a less uncomfortable test tone of 75db. And, all channels are calibrated equally to that 75db level. Our AVR's then do an internal recalculation to add 10dB to the regular channels, so that a master volume of 0.0 will equal approximately 85dB. And, once that internal recalculation takes place, the audio system is now calibrated to Reference (allowing for 105dB max for the regular channels, and 115dB max for the LFE channel) at a master volume of 0.0.
Once an audio system has been calibrated to Reference, whether that particular audio system will subsequently be able to play 105dB peaks in the regular channels and 115dB peaks in the LFE channel is entirely a function of the capabilities of the individual audio system in that particular room. And, as mentioned earlier, whether an individual user will ever decide to play his audio system at Reference levels, even if the system is capable of doing it, is entirely an individual choice. In fact, most people don't ever play their audio systems that loud.
[Based on anecdotal information, the average HT listening level is probably in a range from about -20 to -10 MV, with a master volume of -10 sounding twice as loud as a master volume of -20. (Each additional 10dB of SPL equals a doubling in perceived loudness.) Many people will be either below or above that average range, but most of us probably fall within it.
As noted in Section I, a number of audio experts have determined that rooms under ~20,000^3, which would include almost all home theaters, sound anywhere from +5 to +9dB louder at the same SPL than would be the case in a commercial cinema. Commercial theaters are much larger than home theaters, and higher sound levels don't sound quite as intense in those large spaces. That probably helps to account for the fact that relatively few people, even in treated rooms, listen at 0.0 MV.
Regardless of preferred listening levels, however, it is important to have all of the channels in an HT system playing at the same volume at the MLP, so that sounds from all of the speakers (and subwoofers) will be in proper balance. But, as a practical matter, starting with all of the channels (including the combined sound of all the subwoofers) playing at the same volume is probably also the only way to set the audio system to Audyssey Flat. The intent of the Flat response curve is to have every frequency from down to as low as 10Hz, and as high as about 22KHz, play +/- 3dB. Audyssey attempts to accomplish this by setting control points which add boosts to some frequencies, and cuts to other frequencies, within every channel in an audio system.
Speakers and subwoofers are typically designed to play a reasonably flat frequency response, so that some frequencies don't stand out in comparison to others. But, once those speakers, and especially subwoofers are placed in a room, the room itself will affect the frequency response, causing peaks at some frequencies and dips at others. Proper speaker and subwoofer placement within a room will help, but bass frequencies in particular often need help from some form of room EQ to even-out the frequency response.
Audyssey attempts to provide that help by setting control points within every channel. When an audio system has been EQed to a relatively flat frequency response, the room is at least partly taken out of the equation, allowing the speakers and subwoofers to play a naturally flatter frequency response. This is generally believed to be particularly helpful for bass frequencies (from about 500Hz down), and may also be helpful for higher frequencies, depending on the particular room and listener preferences.
The Audyssey Reference curve changes Flat by creating a slightly downward curve to the high-frequency response. It is the default setting after an Audyssey calibration. The Reference curve slightly rolls-off the treble frequencies above 4,000Hz (by about -2dB) and it adds more roll-off (about -6dB total) above 10KHz. It also adds mid-range compensation (a -3dB dip centered between 2,000Hz and 3,000Hz). Many people prefer those high-frequency roll-offs, and it is strictly a YMMV issue as to which of the two Audyssey settings we use. But, to create either the Flat curve, or the Reference curve, Audyssey needs to start with all channels and frequencies playing at the same volume at the MLP.
[Again, it should be noted that a similar methodology, of setting all channels, including the combined sound of the subwoofers, to the same volume, is used by other systems of automated or manual calibration, whether any room EQ is being attempted or not.]
** Occasionally, someone may say that he prefers to hear the "natural" sound of his speakers, or subwoofers, without any room correction at all. As far as I am concerned that is an entirely personal decision. There isn't any right or wrong way to listen to audio, in my opinion. But, I have never really believed that, given a good calibration, Audyssey is likely to significantly change the "natural" sound of speakers. Horn speakers still sound like horn speakers, and electrostatic panels still have their own characteristic sound. What Audyssey may do, however, is to change the way that speakers and subwoofers interact with a room. After all, changing room/transducer interaction is the purpose of implementing room correction. Whether that change is positive or negative is up to the individual to determine.
What any system of room correction attempts to do is to change the interaction of the speakers with the room, in order to allow the speakers (and subwoofers) to play with less distortion, caused by room-induced peaks and dips in the frequency response. As noted below, those peaks and dips are especially prevalent in bass frequencies. Again, whether Audyssey or any other system of automated room EQ is successful in creating a smoother frequency response, and whether attempting to do that results in improved sound quality, is strictly up to the individual listener to decide.
And, that may depend on variables which include the specific room, and room furnishings and treatments; the speakers/subwoofers, and their specific positioning; the relative care taken during the calibration process; the AVR settings employed; and the personal preferences of the particular individual.
But, in order to fully experience Audyssey, I would encourage users to take pains in system setup and in their Audyssey calibrations. As noted in Section I, differences in system setup can yield different results, as can differences in microphone placement. Some calibrations may sound and/or measure better than others, and once a good calibration routine is developed, it is a good idea to keep a record of the mic placements that produced the most positive results. I would also encourage users to experiment thoroughly with various settings, such as the Audyssey Reference curve versus Audyssey Flat; with DEQ on and off; and of course with bass boosts. Individual setting changes, such as those, can sometimes significantly change the nature of the sound.
Section II-B: Why We Add Bass After Calibrations
The room strongly influences bass response, due to the action of room modes, causing peaks and dips at various frequencies. That is why Audyssey can be so helpful in EQing subwoofers. Audyssey can implement boosts up to +9dB and cuts up to -20dB, at selected frequencies, in all of the channels, in order to achieve a flatter frequency response. When Audyssey is successful at flattening out most of those bass dips and peaks (at least to some extent) the result may be a smoother, clearer, and more uniform sound.
That less distorted and less boomy sound, without some frequencies peaking at much louder SPL's than other frequencies, may give the impression that there is less overall bass playing. And, there may actually be less bass playing, at some frequencies, if a particularly noticeable frequency (say at 50 or 60Hz) were peaking quite loudly prior to EQ. Just hearing all of the bass frequencies, in better equilibrium with each other, may contribute to the initial impression that there is less bass.
But, setting aside whatever impressions of lower bass volumes we may have, when we hear less distorted bass in our rooms, there is more to it than just hearing a smoother frequency response. Most people don't listen at Reference Volumes (0.0 MV) which is where the low-frequency content in 5.1 movies was mixed to be in correct balance with other frequencies. Once the volume level of a movie is reduced below Reference, in a typical home theater, those low-bass frequencies will typically be harder to hear, in relation to the frequencies where our hearing is stronger.
That is because, as the volume level drops, it will appear to drop faster for frequencies outside our normal hearing range. And, that particularly includes low-bass frequencies, which carry much of the special audio effects in movies. So, in a properly calibrated audio system, listening at anything less than very loud volume levels, it is pretty normal to perceive the bass as sounding too soft.
[It is worth noting that people sometimes attempt to double-check their subwoofer volumes, after an Audyssey calibration, with an SPL meter. Even if the SPL meter is correctly set to C-weighted Slow, however, it may not be able to accurately record sound pressure levels below about 40Hz, where Audyssey is measuring and correcting SPL. Unless the SPL meter is a slightly more expensive one, which is calibrated, it may not yield correct readings for subwoofer frequencies. That is particularly true for smartphone apps, and for the inexpensive meters sold at places like Radio Shack. They can be off by as much as -10dB or more at low-frequencies.]
After the level-matching process from mic position 1 is complete, the low frequencies (which, as noted, are harder for us to hear) are playing at the same volume as all of the other frequencies. This phenomenon of lower frequencies being harder to hear than higher ones (except for very high-frequencies) is well documented. Our hearing is strongest from about 500Hz to about 5,000Hz. So, frequencies played by our subwoofers may require more volume than the frequencies played by our regular channels. Some additional explanation of this is included in the section on DEQ, and in the Addendum on the thread history. A more complete discussion of the Equal Loudness Contours, which define our perception of loudness at different frequencies, has also been included in Section VII-C.
[For the sake of this discussion of bass boosts, it should be noted that the Equal Loudness Contours, which are a slight modification of the original Fletcher Munson Curves, are based on averages in normal, healthy human hearing. It may be assumed that, as with other human attributes, hearing generally follows the shape of a bell curve, with some individuals hearing bass frequencies somewhat better, and others hearing those same frequencies somewhat worse than the average. And, our hearing may (will) change somewhat as we age. Therefore, a given individual may be able to hear lower frequencies relatively better, or relatively worse than would be predicted by reference to the Equal Loudness Contours. Consequently, the final bass levels that we pick may be partly a function of our individual hearing capabilities, and also partly a function of our specific psycho-acoustic preferences.]
Although a lot of the discussion so far has focused on movies and on Reference levels, our desire for stronger bass may not be limited to 5.1 movies or TV shows at below Reference levels. Even if we are watching a 5.1 movie at Reference levels, some of us might still prefer to have more bass than what Audyssey provides with a flat bass frequency response. DEQ won't add any bass at all at a listening level of 0.0. The decision of how much bass we want to hear (at any listening level) is an entirely personal one, which may depend on a number of factors.
We may also prefer to apply bass boosts to music. And, we may wish to add sub boosts for some of the same reasons that we would add them for movies. Even in the absence of special effects in movies, we may not hear low-frequencies in music as well as we hear those in our optimal hearing range, or we might just prefer more bass, period. That would be especially true at lower listening levels. And, some music and some TV shows may have less low-bass in the soundtrack to begin with. For instance, some older music may have very little content below 60 or 80Hz, due to the nature of the recording process, and we might be used to hearing more low-bass in our HT systems than that. So, some people might wish to boost the subs just to hear what sounds like a more appropriate amount of bass.
[I think it is important to emphasize that the degree of sub boost selected for any program content is entirely a matter of personal preference, as is the overall listening level. And, individual preferences may change, as we go from one source to another, from one song or movie to another, or depending on our moods, from one day to another.]
Audyssey's DynamicEQ (commonly abbreviated as DEQ) is a separate software program which boosts the low-frequencies in all of the channels, including the .1 subwoofer channel. It also slightly boosts the high-frequencies in the regular channels. It is engaged by default whenever an Audyssey calibration is run. The boosts that DEQ adds are intended to, at least partly, compensate for the inherent difficulty in hearing lower frequencies (and to a lesser extent, high-frequencies) at below Reference volume levels. DEQ is explained in detail in Section V.
How much boost DEQ adds varies depending on the MV selected, with more boost added as listening levels reduced below Reference (0.0 MV) at a rate of about +2.2dB per -5dB below Reference. So, at -15 MV, for instance, DEQ would add a little over 6dB (6.6dB) of bass boost to all of the channels, including the sub channel.
Whether DEQ fully restores bass equilibrium to movie soundtracks is an interesting question. Most people seem to prefer more bass boost than DEQ provides, and typically add an independent sub boost, even with DEQ on. With DEQ engaged, the typical sub boost appears to average about +3dB to +6dB. With DEQ off, sub boosts are typically much larger. Additional information, regarding DEQ, may be found in a later section. But, with or without DEQ, the question of how and where to add a sub boost is important for most people.
Section II-C: Where And How To Add Bass
Most modern commercial subwoofers have a gain (sometimes labeled "volume") control. That gain control helps to determine how much power will go from the sub amp to the driver (woofer). The way it works is that voltage goes from the AVR's sub out amplifier, to the subwoofer amplifier, which then amplifies that signal according to where the gain level of the subwoofer is set.
During the Audyssey calibration, the initial setting of that gain control will determine where Audyssey sets the trim level for the sub(s). So, if the initial gain control is high, Audyssey will set a low trim setting in the AVR (such as -9) in order to insure that the sub is playing 75dB at the MLP, just as all the other channels are. If the gain control setting is low, Audyssey will set a high trim level in the AVR (such as -3.0, or 0.0, or even +3.0) to insure that all of the subwoofers in a system are playing at that same 75dB level. A simple way to think of what happens during the initial calibration is: high gain level = lower trim level; low gain level = higher trim level.
[To emphasize this point more clearly, during the automated calibration process the subwoofer gain level and the AVR trim level are inversely proportional. For every decibel that we increase the gain level on our subs, the AVR will subtract one decibel from our AVR trim level. So, if we start with our subwoofer gain level at X, where X = 75dB, the AVR will set our subwoofer trim level at Y. If we set our subwoofer gain level at X + 5dB (80dB) our AVR will set our trim level at Y - 5.
That is because it is the AVR's job to make all of the channels play that same 75dB volume. Otherwise, our HT system can't actually be calibrated to the Dolby/THX Reference level that was our original target. The same thing would be true, in reverse, if we set our subwoofer gain level below 75dB, to 70dB, for instance. Our AVR would raise the AVR trim level by +5dB to compensate for the low volume of the subwoofer amplifier. Doing that would insure that all of the channels were calibrated to play that same 75dB at the MLP.]
So, that explains the relationship between the gain control and the AVR trim level during the initial calibration. But, what about after the calibration? If we want to add a subwoofer boost after Audyssey has set all of our channels to play at exactly the same volume level, how should we do it? should we use the gain control, or should we use the AVR trim control? We can actually use either one, but the decision of which one to use in a particular situation is just a little complicated.
The first thing to understand is that it is often desirable to make the subwoofer amplifier amplify the signal which is sent to the driver, rather than trying to have too much voltage coming directly from the AVR amp, because the subwoofer amplifier is much more robust and powerful than the amp in the AVR. This can be an extremely important point, because using the subwoofer amp for substantial volume levels may help to prevent clipping of the pre-out signal coming from the AVR. Clipping is a form of distortion, due to an alteration in the waveform. It can be audible in some cases, and if prolonged, can lead to overheating the voice coil in the woofer. When a waveform is clipped, the round top of the wave is squared-off---clipped-off.
That can happen when a subwoofer attempts to play low-frequencies with too much voltage. Simply raising the trim control in the AVR may not result in sufficient clean voltage going from the AVR pre-out to the subwoofer, if the subwoofer gain is set too low. Depending on the AVR, the pre-out signal may be slightly higher or lower, but in all cases that pre-out signal will not be as robust as what the subwoofer amplifier can produce. So, turning-up the trim may not actually result in more undistorted voltage going to the subwoofer. Instead, the AVR may send out a clipped signal at higher trim levels. That's why you typically want to avoid having too much voltage coming from the AVR, and insufficient amplification of the signal, occurring in the subwoofer amplifier.
It is necessary to have enough voltage coming from the AVR amp to turn-on the subwoofer from its Auto-On mode, and that can vary a little bit among some older AVR's. The -5 AVR trim number is sort of an arbitrary number, but several subwoofer makers suggest using that as about the max AVR trim setting for subwoofers. Holding our AVR trim at about -5, or lower, forces us to use the gain controls on the subs to add any really substantial subwoofer boosts.
After running Audyssey, simply making any adjustments in sub boost using the gain control on the sub(s) would insure that the sub amp is being used. So, that would be a perfectly good way to add sub boost. And, as noted a little further down, using a higher gain control may enable some subwoofers to achieve higher SPL's than they can with low gain levels and high trim levels.
But, most people find it more convenient to make adjustments using the AVR trim controls, with a remote control. And, in that case, it is desirable to start with a high sub gain level, and a low AVR trim level. Remember that a high gain level = a low AVR trim level. So, we would need to take certain steps during the calibration process if we wanted to have a low AVR trim level--let's say in about the -10 range. And then, we could adjust the trim level upward after the calibration. Using the trim settings in the AVR to make sub volume adjustments, after running Audyssey, allows the user to make convenient and fairly exact (.5dB increments) adjustments to subwoofer volume, by using the AVR remote.
Typically, in order to achieve a low AVR trim level, though, it will be necessary to start with a measured sub SPL of higher than 75dB. An SPL level of about 78dB to 80dB may be required. That would be in the red zone for Denon/Marantz units during the subwoofer level-matching process. Audyssey is specifically trying to set the sub(s) SPL to 75dB. That is in the green zone on Denon/Marantz. However, to get a strongly negative trim level, a higher than 75dB level will be required, and that will be in the red zone. The specific SPL used is not as important as the resulting low AVR trim level.
It should be noted that there is no harm in telling Audyssey to proceed with the calibration, even though the subwoofer is not in the green zone. That notification of red zone just gives owners the opportunity to adjust the gain on the externally-powered subwoofer to the same 75dB volume level which Audyssey is trying to achieve for all of the channels. But, if the owner chooses to proceed with the calibration anyway, Audyssey will simply calibrate the subwoofer(s) with a low trim level, and that is typically exactly what we want it to do. That way, we can start with a low AVR trim setting, and add some subwoofer boost in a very convenient way, while remaining at about -5 or lower with our AVR trim.
It should be emphasized that there is no particular reason not to just use the gain control on a sub to add volume post-calibration. For people wanting to add really substantial bass boosts--up to, or in excess of 10dB or 12dB, some gain increase, in excess of the original gain setting, is generally necessary anyway, in order to achieve the bass boost desired by the user.
* [It should be noted that some Denon AVR's have a feature called "Subwoofer Level Adjust". When this feature is used, the subwoofer trim level is reset to 0.0, and the adjustment is made on top of that. So, starting at -11.5, post-calibration, and adding 5dB of boost with that feature, will actually result in a net trim level of +5.0 in trim, instead of -6.5 in trim. That is a net increase of +16.5dB instead of +5dB. It is highly recommended to turn that feature off, and instead to make any necessary subwoofer volume adjustments either with the Channel Level Adjust in the Audio menu, or with the trim controls in the Speaker: Manual: Test Tone area of the Denon AVR.
Edit: Apparently, this glitch has been fixed in newer Denon models, and with recent Denon firmware updates. If your AVR does not have an on/off switch for the Subwoofer Level Adjust, using that feature will add the correct amount of subwoofer volume, just as would be the case if you used the test tones. If your AVR does have the on/off switch, you can test the SLA feature to determine whether it is adding the appropriate amount of subwoofer boost.]
[It should also be noted that ARC Genesis doesn't allow users to ignore the gain setting zone in the way that Audyssey does. So, using high gain settings to achieve low-trim settings, doesn't work well with the most recent version of ARC. When using ARC Genesis, users should simply follow ARC's instructions during the Quick Measure process. After ARC has run its automated calibration routine, users can still turn-up the gains (symmetrically) on their subs, if necessary, in order to achieve more bass, or to reduce high AVP trim settings. It is always going to be a matter of listener discretion as to whether or not to avoid high AVR/AVP trim levels. As with any subwoofer/AVR combination, there may or may not be audible clipping with high AVR trim levels.]
To continue the general discussion of where to add subwoofer boosts, the usual recommendation to employ the AVR trim is more a matter of convenience and of accuracy than one of necessity. Some subs don't have digital gain controls, for instance, so fine-tuning the gain can be more difficult, as can on-the-fly adjustments during a particular movie, or music listening session. And, it gets even less convenient when multiple subs are connected together, or when gain controls are difficult to access easily. Using the trim controls in an AVR allows for very convenient and precise adjustments in sub volume. But, the most important thing is to make sure that the real boost comes from the subwoofer amp, and not just from the AVR, whichever adjustment method is ultimately employed.
Assuming that some initial sub boosts are to be accomplished using AVR trim, then starting with a low trim level post-calibration would be helpful. A low trim level might be defined as -9 to -11, but not exceeding -11.5 in Denon/Marantz units. (With other manufacturers, just determine what the minimum trim level settings are in order to ascertain what your optimum low trim setting should be.) As stated earlier, it may take an SPL of 78dB, or higher, to achieve that optimum low trim level. However, it is important not to go lower than -11.5 in trim, in Denon/Marantz units. (For example, I believe that the lowest trim levels we should calibrate our audio systems to are -14.5 with Onkyo, and -9.5 with Yamaha, based on their respective trim level limits of -15 and -10.)
If a trim level of -12 is set, with Denon/Marantz units, there is no knowing what the actual volume of the subwoofer is. The AVR simply ran out of negative trim at -12. The actual sub volume might be 80dB, or even 85dB. If so, you might not like the way it sounds to have your sub so much louder than the rest of your system. And, you would not have an easy way to turn down the subwoofer volume, if your trim level were already at the lowest setting. You also could never be sure what your actual sub volume is, and as a result, you could find yourself running out of headroom sooner than expected. So, for instance, you want a negative trim setting not exceeding -11.5 in Denon/Marantz units.
Think of the process of adding a sub boost this way. When you raise the gain level in the sub, so that the sub produces more than 75dB at the MLP, you are making a deposit in the bank, of amplifier power from the sub. So, for instance, let's say you start with a trim level in the AVR of about -9 to -11. Now, you can withdraw amp power from the bank, using your AVR trim control. You would, for instance, do that by increasing your trim setting to about -6 or -5. As noted earlier, a +3 to +6dB boost would be pretty typical, even with DEQ engaged. But, there is no free lunch. As you begin to approach 0.0, the bank deposit of amp power that you made with the higher gain setting is used up, and now you are using AVR amp power, which as noted, is not as powerful. Using AVR amp power can, in some instances, result in clipping (distorting) your subwoofer(s) or it can, in some cases, result in undesirable mechanical noises.
[Listeners sometimes mention the importance of using only the AVR trim to add subwoofer boosts, because they are able to know exactly how much additional SPL they have added that way. I have never been convinced of the importance of that. As a general rule, we are simply adjusting our subwoofer volume to match our personal listening preferences, anyway.
If we suspect that we may be running out of headroom, that is a separate question, irrespective of the precise amount of subwoofer boost we are adding. If we have ample headroom, then I'm not sure that the precise amount of SPL we have added is really very important. Some subwoofers add about 1.5dB per click, where there are detents in the analogue dial. Others may add about +3dB. The subwoofer maker can usually confirm the amount per click if listeners are really interested in knowing approximately how much SPL each click represents. And, of course, listeners can always measure their subwoofer SPL, as they add gain, if they want to know more precisely the amount of subwoofer boost added post-calibration.
Some listeners also wonder how important it is to keep track of how many clicks they have added post-calibration, and are concerned about not being able to do that with analogue dials which don't have sufficiently fine detent markings on the outside of the dial. I know that some listeners have used washable magic markers, or small pieces of tape, to mark the original post-calibration setting, before they add any additional gain boost. We can be pretty creative about that kind of thing if we are really curious.]
Section II-D: Master Volume Levels And Sub Boosts
There is a relationship between subwoofer volume and master volume (MV). As your MV increases, the subwoofer volume goes up correspondingly, and more demands are placed on the sub. It is important to remember that the subwoofer is not only playing the LFE channel, but also providing bass support for all of the other channels in a typical HT system. So, as the MV increases, the demands on the sub go up much faster than for the other channels, particularly in a movie with a lot of low-frequency content. It is worth noting that 5.1 movies (and some bass-enhanced music) can have very low frequency content in all of the channels, and not just in the LFE (low frequency effects) channel. The subwoofer has to (and should) play all of that low frequency content.
It is recommended by a number of subwoofer experts, two of whom are quoted in the FAQ, that it is advisable to keep sub trims well in the negative range (below 0.0). That is particularly important as MV's approach, or exceed, -10. In Denon/Marantz units, that is 10dB below Reference (or 70 on the absolute scale) in your AVR master volume. Both of those experts quoted in the FAQ, Ed Mullen of SVS, and Mark Seaton have, subsequent to the entries in the FAQ, recommended staying well in negative trim levels, period. To follow their advice, and to avoid the possibility of distortion, we would want to keep our trim levels in about the -5 range, or lower, at even moderate listening levels. Again, that is easy to do by simply raising the gain on the subwoofer(s).
* [In addition to the possibility of clipping the subwoofer signal, with higher AVR trim levels, there is another potential reason for keeping gain levels relatively high. It is addressed in the following section titled Gain Settings and Maximum Sub Output.]
High sub gain levels, which still result in high trim levels, are indicative of a sub which is under-powered for the space, and/or the distance from the MLP. It could also be indicative of a specific placement problem, where either the sub or the MLP is located in a null. In the first instance, the only remedy would be a more powerful sub, or multiple subs, or a different (probably closer) sub placement. In the situation where the sub or the MLP were located in a null, a subwoofer crawl should be done to determine proper sub placement. Although subwoofer placement is not a direct part of this particular discussion, it is a very important factor in sub performance.
If you never intend to approach about -15 MV or higher, then the advice to set your sub gain high enough to obtain a strongly negative trim level might be less important. (Even then, however, that could be somewhat dependent on the use of DEQ and/or the use of independent sub boosts.) And, if you don't believe that you will ever want to listen at high volumes, or to boost your subs, then starting with a trim level of about -5 or -6, should be perfectly fine.
But, most people on this and other threads seem to average at least a +3 to +6dB bass boost after calibration, and some people add much more than that. When DEQ (with its own bass boost) is not employed, boosts of +12dB, or even more, are not uncommon. So, the advice you will most commonly see on this thread is to start with a negative trim setting of about -9 to -11 post-calibration, in order to maximize your ability to add sub boost, with your AVR trim control, while still using the sub gain you deposited in the bank.
Although this advice is not consistent with the explanations and recommendations in the FAQ, the more current advice supersedes the older advice in the FAQ. I would personally recommend following the advice to maintain a negative sub trim, preferably of -5 or lower, as a matter of best practice, even if you believe that you will never approach -10 or -15 MV. (As noted in the section just below, keeping a relatively higher gain level may work to your advantage, in any event.)
There is no telling who might, inadvertently or otherwise, run the volume control up on your system, or when unexpected peaks in very low bass (in electronically-enhanced music, or in movies) might cause some distortion to occur. And, if your sub happens to be approaching its max output limits, even at lower master volumes, the lower trim level would provide an additional measure of confidence that you weren't clipping the subwoofer signal.
While it is unlikely that most good modern subs would be damaged by a bit of distortion, or by an inappropriate use of AVR amp power, I know of two well-documented instances of a JTR Orbit Shifter, which is an extremely powerful and well-made sub, frying a voice coil (due to overheating) just from playing electronic music, downloaded from YouTube, at a very high volume, with a high AVR trim level. (Both instances involved the same user, who didn't quite remember the lesson from the first instance.)
And, even if no damage could ever be done as a result of clipping the sub signal, listening to distorted bass is sort of antithetical to the whole idea of good sound quality, and of using automated room EQ to achieve it. Clipping also consumes another +3dB of amplifier power. Where each +3dB of SPL equals a doubling in amplifier power, that +3dB increase, due to clipping, is significant. When proper gain/trim protocols are followed, it is also less likely that inappropriate noises, such as port chuffing, or of drivers hitting limiters, could occur prematurely. So, an ounce of prevention is worth more than a pound of cure, in this case.
Again, you can use a combination of increased subwoofer gain, and some increase in AVR trim, to raise the volume level on your sub to any level you choose, while still maintaining an AVR trim of about -5, or less. (Since originally writing this guide, I have seen even more recommendations from subwoofer makers to be at -5 or less in the AVR trim.) That will help you to avoid the possibility of clipping the subwoofer signal. And, raising the gain control on the sub(s) post calibration, will have no effect at all on the way that Audyssey EQed your system.
Section II-E: Gain Settings And Maximum Sub Output
There is another aspect to the gain/trim issue that is worth mentioning. Depending on how the DSP in a given subwoofer is implemented, the subwoofer may only be able to achieve max output levels with the gain control set very high. Some subwoofers are only able to achieve maximum output levels when the gain control is set to, or very near, the highest setting. So setting a lower gain control, and a correspondingly higher AVR trim control, might not result in the same amount of peak bass SPL, irrespective of issues of clipping.
Apparently, this issue may be more common in subwoofers with digital (rather than analogue) controls. But, according to several examples I have observed from various threads, the issue is not at all limited to subwoofers with digital controls. Some subwoofers with analogue controls may have the same issue of not being able to achieve higher max output levels with low gain settings.
How important this max output issue actually is probably depends on the situation. For instance, I believe that a relatively lower gain setting might cause a ported subwoofer to chuff prematurely. Again, depending on the situation, even someone who is listening at a fairly moderate listening level, let's say -15 or -20 MV, might experience issues if he were using a significant subwoofer boost, either independently or on top of DEQ.
Putting a sudden peak demand on the subwoofer, with the right low-frequency content, might not enable the subwoofer to access the full output that it is designed to produce, if the gain level isn't fairly high. In that case, the subwoofer just wouldn't play the low-frequency content at the volume it was supposed to. In other words, it would simply stop getting any louder during that peak content. Whether we would even notice that, or whether we would hear the subwoofer make any audible sounds of distress, are separate questions.
But, unless we are sure that our subwoofers can achieve max volume levels with low gain settings, it is probably a good precaution to keep gain levels fairly high. Typically, that will mean using corresponding lower AVR trim levels for our subwoofers. This is not an issue that I have often heard addressed by subwoofer makers, but I suspect that many would intuitively know that some subs produce max volumes only with high gain levels.
This is just speculation on my part, but I think one reason that this issue isn't discussed more is because sub makers are not typically testing their subs as part of a calibrated HT system. So, they aren't dealing with gain/trim relationships at all in their design and testing process. When they want to push one of their subwoofers to its limits, or they want to measure its maximum output, they just max out the gain control on the subwoofer itself. It is only when subwoofers are calibrated as part of an HT or audio system, with an inverse relationship between gain and trim, that this becomes an issue. But, I believe that it can be an issue, and I believe that is another potential reason for attempting to keep gain settings fairly high.
CEA 2010 testing, performed by Data-Bass, always measures max output with gain controls at the maximum setting. And, as noted, some subwoofers may not be able to produce those same max SPL numbers, that we see on Data-Bass, or from other professional sources, with lower gain settings. This won't be true for all subwoofers, but as a matter of best practice, I believe that it may be generally advisable to keep gain settings fairly high, and AVR trim settings fairly low, in order to maximize available headroom. An exception to this general policy could be a situation where a lower trim setting didn't successfully power a subwoofer on, when it was set to Auto On mode. But, that would be extremely unusual with most receivers and processors.
* People with some Yamaha AVR's are apparently much more likely to experience issues with subwoofers not turning on automatically unless AVR sub trim levels are relatively high--perhaps even fairly close to 0.0. That is due to the lower voltage signal sent from some Yamaha AVR's to the subwoofer. In some cases, this may have been due to some defective sub outs on some Yamaha AVR's, which were replaceable under the Yamaha warranty. Yamaha AVR's from about 2017 on were previously reported to have addressed the problem, but that doesn't seem to always be the case. Both Yamaha and Onkyo AVR's may also experience a calibration (level-matching) problem due to low voltage signals to the subwoofers. That issue is addressed in the last few paragraphs of this section.
If subwoofers will not turn on automatically in Auto mode, without higher AVR trim levels, then the higher trim levels may be slightly less likely to lead to clipping issues, since the voltage from the AVR was lower to start with. Some Yamaha owners use a Y-connector into both subwoofer inputs in order to double the voltage going to the sub. And, that typically resolves the Auto On issue. Of course, Yamaha owners can also choose to just leave their subs on all the time, if the Auto On issue proves to be a real problem. That will consume slightly more energy, but will not affect the operation or longevity of the subwoofer.
AVS member @Basshead recently mentioned a clever solution for achieving more headroom, with higher gains and lower trim levels, which seems to circumvent the Auto On problem with Yamaha AVR's. He went from a -1.5 subwoofer trim level to a -4.5 trim level, with a comparable gain boost, and obtained +3dB more headroom, prior to clipping. But, his subwoofer didn't power-on reliably when watching TV at very low master volume levels. So, he lowered the trim levels on all of his other channels by -3dB, and raised his MV level by +3dB, and is now able to have his sub power-on reliably for low-volume TV content, while still having more headroom available for louder movie viewing. This is an additional technique that Yamaha owners might wish to try.
One final issue involves Yamaha AVR's which yield abnormally high trim levels no matter how high the subwoofer gain levels are turned-up. This can be a much more significant problem than the auto-on issue. In a recent example from early October of 2021, a new Yamaha RX-A8A exhibited this issue of setting abnormally high trim levels with the subwoofer gains also set high. Some before-and-after screen shots of trim levels, showing the problem being solved with Y-connectors, are illustrated on Page 218 of the Guide thread. If you believe that your Yamaha (or Onkyo) AVR may be exhibiting similar behavior, it could be worthwhile to look at the pictures on Page 218.
Apparently, this can also be an issue with Onkyo AVR's. The typical voltage sent from AVR's to subwoofers is about 2.0V or slightly higher. I'm not sure what the voltage coming from some Yamaha AVR's is, but I have been informed by AVS member @fattire that with some Onkyo's it is only 0.9V. If the subwoofer receives substantially less voltage than the typical 2.0V, then its performance may be adversely affected, and that limitation may show-up during calibration. The subwoofer trim levels may be set too high even with very high subwoofer gain levels. That can be an indication that the subwoofer is not able to reach its normal output potential.
Where this problem is believed to be occurring, the way to correct it is to use a Y-connector into both sub inputs. That will double the voltage going from the AVR to the subwoofer. So, using Y-connectors can allow subwoofers to turn on automatically in some cases, and they can also be used to allow the subwoofer to achieve its full operating performance. Once again though, the Y-connectors are only effective where the voltage coming from the AVR is insufficient. They won't improve on the inherent performance capabilities of the subwoofer. Its own amplifier will determine the subwoofer's inherent capability.
This is an example of the type of Y-connector which would be used to increase the voltage going from the AVR to the subwoofer:
Section III: Setting Crossovers:
This general discussion of bass-management, and of setting crossovers, applies to other systems of automated calibration and not just to Audyssey. The same questions come up so many times that I think it is worth emphasizing some of the basic crossover concepts. We use crossovers between our speakers and our subwoofer(s) in order to bass-manage our audio systems. In audio systems where there is no subwoofer, there will be no bass-management required, and speakers will always be set to Large or Full-Range.
The Subsections in Section III are as follows:
III-A: Crossovers from Speakers to Subwoofers
III-B: Low Frequency Effects Channel
III-C: Cascading Crossovers
III-D: Bass Localization
III-E LFE+Main
Section III-A: Crossovers From Speakers to Subwoofers:
Subwoofers can be used with two speakers in a stereo system, or they can be used with a 5.1, or larger, audio system. Whenever they are used, it is necessary to have a way to cross over from the speakers to the subwoofers, so that the subwoofers can play bass content below a designated frequency. Good subwoofers are designed for the sole purpose of playing bass frequencies below about 150Hz.
(Subwoofers will typically have a setting labelled "LPF" (low-pass filter) or "Crossover". It may be an analogue knob on the amplifier plate. That setting should generally be at the highest setting, which will usually be 150Hz. That will allow the subwoofer to play frequencies up to 150Hz, before it starts to roll-off. An exception to that general rule is explained in Section III-C: Cascading Crossovers.)
As frequencies go below about 120Hz, and especially below 80Hz, subwoofers typically perform their specialized function much better than even large tower speakers can. We set crossovers to allow our subwoofers to take over duties below a selected frequency. That selected frequency will depend to some extent on the native capability of our speakers, and it will depend somewhat on the speakers' placement in the room, since room placement will strongly affect low-frequency performance of any transducers in both positive and negative ways, as explained in Section I.
Section I also noted the importance of placing tower or bookshelf speakers where they can point toward the listener, so that mid and high-frequencies will be heard clearly. The advantages of spreading speakers apart to achieve a wider soundstage, and potential issues with early reflections from side-walls were also discussed. But, those are all issues affecting mid and high-frequencies. And, they may necessarily dictate the placement of our front speakers, if we want to maximize our sound quality. The room geometry and the furniture arrangement in the room, may also be factors in the positioning of our front speakers.
Bass frequencies have different issues with respect to placement though. So even if our front speakers are not optimally positioned with respect to bass frequencies, our subwoofer(s) may be able to compensate for that. Good subwoofer placement can be completely independent of the placement requirements for our tower and bookshelf speakers. This is why, even those of us with very capable speakers, may wish to have subwoofers which we try to strategically locate for optimal bass performance, and which we bass-manage for optimal integration with our speakers. That is especially helpful for movies, where the low-bass demands can be very significant.
When we look at a 5.1, or larger, system, we see even more importance attached to the subwoofers. First, the subwoofers must provide bass support for all of the regular channels, just as they would in a two-channel system. In 5.1 content, those regular channels can have bass peaks up to 105dB, depending on master volume and bass levels, and the frequencies can go very low at times. Second, the subwoofers must play all of the content in the LFE (low frequency effects) channel, at peaks of up to 115dB, also with potentially very low-frequencies. Subwoofers which are powerful enough for the room, and for the individual listener's preferences, perform this double duty just as they are designed to.
Those two different subwoofer functions are controlled by two different mechanisms. As explained below, the bass content redirected from the speakers is controlled by crossovers. The LFE content is controlled by a low pass filter, called the LPF of LFE. That LPF setting is explained in more detail in Section III-B. The default LPF setting in most AVR's is 120Hz.
The low-bass content in the regular channels is controlled by crossovers set for each pair of speakers (or for the center channel). In order for bass to be redirected from the regular channels to the subwoofer(s) speakers must be set to Small, with a crossover. The assignment of crossovers for each channel is accomplished during the initial calibration process, and then may be modified by the user. Crossovers may be assigned globally (such as 80Hz for all channels) or they may be assignable individually, for speakers pairs and for the center channel, depending on the specific AVR.
As stated earlier, where we have subwoofers in our audio systems, and wish to employ them as they were designed to be used, some system of bass-management is necessary to split the signal between the speakers and the subwoofer, so that the subwoofer can handle low-frequency content in the regular channels, while the speakers continue to play all of the other content. And, that split can only be accomplished through a setting of Small, with a crossover. Determining where that split between the speaker and the subwoofer should occur starts with the calibration process, where initial crossovers are assigned by the AVR. And, it continues after the calibration process, as listeners adjust their crossovers to achieve their specific listening objectives.
The frequency at which a signal split should occur may be different for different speakers in our audio system, depending on their low-bass capabilities and on their placement in the room. The subsection on room gain (in Section VII-B) explains something of the unpredictable ways that a room, and placement within a room, can affect the bass response of a subwoofer. The same explanation applies to speakers. Room placement can affect a speaker's bass response, which in turn determines where a crossover will be set. (Where crossovers must be set globally for all of the speakers in an audio system, as with Yamaha AVR's, some compromise may be necessary.)
It was noted in an earlier section that distances (timing) and channel trim levels are determined by microphone position number 1. That is not the case with crossovers. Crossovers are set based on the fuzzy-weighted average of the frequency response from all six or eight microphone positions. As a general rule, crossovers in a full calibration will not vary much from where they would be set based on the first few mic positions. But, they may vary slightly, once Audyssey has calculated the FR at all available mic positions.
Audyssey (and other systems of automated calibration) accomplish bass-management during the initial calibration. When Audyssey measures all of the speakers in an audio system, it reports the measured F3 point of each channel to the AVR or AVP. (The F3 point is the frequency where a speaker is reaching the bottom of its frequency response, and rolling-off in SPL by 3dB.) That point at which a speaker begins to roll-off naturally by -3dB will be dictated by both the inherent capability of the speaker, and by its position within the room.
Once Audyssey has completed its measurements of frequency response for each speaker, the AVR then sets that speaker, or speaker pair, to either Small or Large (also called Full-Range), based on its own internal programming. If a speaker begins to roll-off in the mid to high 30's (or lower) the speaker will be set to Large. At any frequency above about the high 30's, the speaker will be set to Small, and a crossover will be assigned. The weaker of two speakers in a pair will control the crossover, as Audyssey is specifically designed not to EQ below the F3 point of any speaker, or subwoofer. Speaker location, with respect to boundary walls and room modes, may make one speaker in a pair roll-off earlier than the other one.
As stated, if a speaker's F3 point is somewhere in about the upper 30's, the AVR will round-up, and set that speaker's crossover to Small with a 40Hz crossover. Crossovers will always round upward, so an F3 point of about 42 or 44Hz would round-up to a 60Hz crossover, and an F3 point a little above 60Hz would round-up to an 80Hz crossover. The exact number used to round upward would probably vary somewhat among different AVR makers, but the basic principle involved is applied in both Audyssey and non-Audyssey systems. From the user's standpoint it is important to note that, without independent measurement, there is no way of knowing exactly where a speaker's roll-off actually occurred. So, it may be advisable to be conservative with crossover settings, after a calibration.
[Sometimes people observe changes in crossover settings that seem to coincide with a change in the type or number of subwoofers. Crossovers may vary slightly from one calibration to the next, and certainly can change due to relatively small shifts in speaker positioning. As explained in the later section on room gain, boundary gain due to proximity to walls can affect the low-frequency response that Audyssey is measuring, as can specific room modes. Small changes in microphone positioning between calibrations can also affect crossovers, as Audyssey is averaging the results of all of the mic positions performed in a calibration. But, Audyssey is only measuring each speaker in isolation, without reference to subwoofers. Any change that appears to coincide with some change to a subwoofer is entirely coincidental.]
The initial setting of Large, or of Small with a 40Hz, or 80Hz, or higher crossover, does not constitute a recommendation, either by Audyssey or by the AVR. This is an important point to understand. The initial crossover setting is simply a modest setting designed to somewhat protect the speaker, while providing information about that speaker (or speaker pair) to the user. The information the initial crossover setting provides tells the user something about where a particular speaker is actually rolling-off by -3dB, at that particular position in the room, and informs the user that no EQ is being performed below that approximate point defined by the crossover. It is then the user's responsibility to interpret that information, and to decide whether to leave the crossover at that initial setting, or to change it.
* It has already been noted that setting speakers to Small with a crossover is the appropriate way to direct lower bass frequencies from the regular channels to the subwoofer(s). As a general rule, I would suggest that where a calibration sets crossovers to 40Hz, an increase to at least 60Hz, or perhaps even to 80Hz, might be advantageous. I believe that the 40Hz crossover setting covers a very narrow range of frequencies from about 36-38Hz to about 42-44Hz. Anything higher than about 42Hz or 44Hz will probably automatically round-up to a 60Hz crossover. With speakers already rolling-off at about 40Hz or so, a good subwoofer should handle that 40Hz frequency, and those at least a half-octave higher, much more effectively. I believe that 80Hz would typically be a much more conservative setting for most speakers.
This might be a good opportunity to give a practical explanation for why it can be advisable to raise crossovers after the calibration process is complete. Let's take the example cited above, of an initial crossover setting of 40Hz. If we see a crossover setting of 40Hz, that may confirm our belief that a speaker (or speaker pair) is pretty capable. But, assuming that a speaker is already down in volume by -3dB, at about 40Hz, what does that really mean in practical terms? It means that at 75dB, the speaker is already running out of gas at about 40Hz. And, 75dB isn't very loud, for bass frequencies, in 5.1 content.
If someone is listening at a master volume (MV) of -20, in a calibrated HT system, that speaker will need to be able to play peak volumes of 85dB with 5.1 content. At -15 MV, that speaker will need to be able to play 90dB for peak volumes, with 5.1 content, and so on as the volume level increases. We already know from previous sections that we can't hear lower bass frequencies as well as those in our normal hearing range, and at -15 MV, with a 40Hz crossover, 40Hz is going to be playing about-18dB softer than it should be at a master volume of -15.
Asking it to play frequencies that it really can't play effectively shouldn't hurt it. That's why the high-pass filter in the crossover makes it play softer once it gets to that F3 point. But, at best we simply won't hear the 40Hz frequencies, and at worst we may hear some distortion, compression, or clipping from that speaker. Alternatively, if we had crossed that speaker at 80Hz, instead of at 40Hz, the subwoofer(s) would have been able to play the 40Hz, and 50Hz, and 60Hz frequencies, much more easily and with much less potential distortion or compression. At a minimum, the sound should be more balanced and the lower frequency sound quality might be clearer too.
Here is another factor we should also consider, if we have something such as Audyssey's DEQ, which is pre-programmed system of loudness compensation. YPAO (on Yamaha AVR's) has something similar to DEQ. DEQ is explained in detail in Section V-A, but briefly, DEQ boosts the bass in all of the channels as listening levels drop below the Reference volume of 0.0. DEQ adds approximately +1.1dB, for frequencies between 70Hz and 120Hz, for each -5dB decrease in master volume. And, it adds progressively more bass boost below 70Hz to a total of +2.2dB at 30Hz and below. At 40Hz, DEQ would be adding about +2dB to each channel for every -5 MV.
So, using our previous example of -15 MV, DEQ would add another +6dB at 40Hz. And, at -20 MV, DEQ would add another +8dB to a speaker which was already trying to play 13dB louder than it actually could. Understanding how DEQ can put additional demands on the low-frequency performance of our speakers may help to explain why some people feel that DEQ makes their audio systems sound bloated. When we talk about compression of bass frequencies, we mean that the lowest frequencies stop getting any louder, while the mid-bass frequencies start to dominate more. That could result in what is often described as boomy or "one-note bass".
None of this means that we shouldn't use DEQ, or that we can't enjoy our speakers to the fullest extent. But, it does mean that we need to understand what an initial crossover setting actually tells us, so that we can help our speakers to play with as little strain as possible. And, that's why most of us added subwoofers to begin with. We wanted to be able to play even lower frequencies, with more volume, than we could obtain from our other speakers. Letting the subwoofers do what they are designed to do may result in better overall sound quality.
Of course, bass localization can be an issue with higher crossovers, depending on where a subwoofer is located, so we may have to balance our interests. A technique that makes subwoofers roll-off faster, above crossovers, is called cascading crossovers. That can help with bass localization, and it may also help with overall bass clarity. Cascading crossovers are explained in Section III-C. And, bass localization is described in some detail in Section III-D.
[FWIW, as noted previously, I think that an initial crossover setting of 60Hz may be especially problematical. Assuming that our AVR is rounding-up from anywhere in the low-40's to about 60Hz or so, then it can be very difficult to know where the F3 point of our speaker actually is. A speaker might be starting to roll-off at about 42 or 44Hz, or it might already be down -3dB in SPL somewhere in the mid to upper-50's. In the absence of measurements to tell me where my actual F3 point is, with a 60Hz crossover setting, or a strong preference for the sound with it set that low, I would personally be more comfortable raising the crossover to at least 80Hz.]
With crossovers already set to 80Hz during the calibration, users might also wish to experiment with slightly higher crossovers, or they might just leave them set at 80Hz. The lower bass frequencies put more demand on a speaker, or a subwoofer, than the mid-bass frequencies do. And, that demand can create distortion in our speakers, particularly at higher volume levels.
I have tried to think of a good rule-of-thumb to use when trying to decide whether or not to be more conservative with our crossovers. I have already recommended increasing crossovers for 40Hz and 60Hz initial settings. It seems to me that master volume levels of about -15, for 5.1 movies and bass-heavy music, would be a pretty safe dividing line for an 80Hz crossover. People listening at about -15, would probably be okay leaving an 80Hz crossover, which was set by the AVR, at that 80Hz frequency. (Using DEQ might also work better with crossovers of at least 80Hz, as DEQ only boosts the frequencies between 70Hz and 120Hz by approximately +1.1dB per -5 MV.)
People listening above about -12 MV, though, might want to raise the crossover slightly higher than that initial setting of 80Hz to relieve any potential strain on the speakers. Remember that bass frequencies will consume more of a speaker's total headroom than higher frequencies will, so the louder we play, the more conservative we might want to be with our crossovers. An exception might involve speakers with very high sensitivity ratings. (Spoiler alert: some manufacturer's may inflate those sensitivity ratings, partly by not showing at what distance the measurement is taken.)
I think it typically makes sense to be a little more conservative, with almost any of our speakers, below about 60Hz or 80Hz. Different speakers, different rooms, and different speaker/subwoofer interactions, however, could all influence the selection of a crossover. Actual experimentation, with careful listening and/or measurement, may be required to make final setting decisions, and the selections may depend heavily on individual user preference. I think that most of us will hear it if our speakers are consistently having trouble playing the material we like, at the volumes we enjoy. In that case, higher crossovers, or lower listening levels, may help to reduce that slightly audible distortion.
Again, it's one thing for a speaker to be able to hit 75dB, at let's say 40Hz, without audible distortion or clipping. It may be a very different thing if that same speaker attempts to hit that 40Hz frequency at 85dB or 100dB. This is why it may make sense to be a little bit conservative with our crossovers.
**
[Listeners who are curious about the specific capabilities of their speakers may want to investigate on their own, via measurements. But, in the context of setting crossovers, this may actually become a little more complex than it seems. Although a listener might choose to try to measure his speakers with a test disk and an uncalibrated SPL meter (such as a Radio Shack meter), I don't think that the results would be very reliable, compared to what the calibrated microphone of an AVR can already do. Uncalibrated SPL meters are notoriously unreliable for lower bass frequencies.
I think that really accurate results would probably require the use of a calibrated UMIK-1 and REW, or some comparable measurement system. The tester would want to see where his speaker was rolling-off by 3db, via a frequency response graph, with the volume at sufficiently loud levels. Or, he could do a compression test, to determine the same thing. If he were using DEQ, he should have it on for these tests. And, he could listen to the audible results during those tests, to identify distortion, clipping, or compression. He could then correlate the results to his typical listening levels.
The following link to the REW thread will help users, so inclined, to understand what is involved in the use of REW:
Simplified REW Setup and Use (USB Mic & HDMI...
For most HT owners, relying on the AVR to have correctly identified the speakers' roll-off points, via the initial crossover setting which takes that roll-off into account, is probably going to be sufficient. Then, understanding that the AVR has identified the roll-off points for us, and has set crossovers accordingly, we can exercise independent judgment on whether to leave the crossover setting where it is, or to raise it. The entire purpose of this subsection is to assist in understanding what the AVR is actually doing, when it sets crossovers, and to assist in the subsequent application of independent judgment in deciding what to do with those initial settings.]
As an aside, the reason that the Large or Full-Range setting is still necessary in modern audio systems is because not every system has a subwoofer, and not every user employs subwoofers for all listening material. For instance, some listeners may choose to listen to a music genre (which may have relatively little low-bass content) with their speakers set to Large, and without any subwoofers engaged. But, in order to employ a subwoofer for anything except the LFE channel, it is first necessary to set speakers to Small with a crossover.
[Some AVR's also have a feature which allows a Large setting with subs employed. The setting is called LFE+Main, or Double-Bass, and is explained in some detail at the bottom of this section, in Section VII-E. That setting may increase the apparent quantity of bass, but may also introduce considerable distortion in the process. It is not generally recommended by Audyssey and others, from an audio quality standpoint, although that is strictly a user-preference issue.]
It should be noted that owners are often surprised by the crossovers set by their AVR's. Sometimes, they are surprised that the crossovers are set so high, and sometimes, they are surprised that the crossovers are set so low, because in either case the crossovers don't align with their expectations. Two major factors contribute to that surprise. First, speaker makers frequently inflate the low-frequency specifications of their speakers. Second, room placement plays just as important a role in the low-frequency performance of our speakers as it does for our subs. The optimum location for a particular speaker (or speaker pair) may give it extra low-frequency response, due to boundary gain or due to favorable room modes at particular frequencies, or it may rob it of some low-frequency performance. Audyssey and other systems of automated calibration and room EQ will simply measure what they detect, and report that to the AVR, which will set the crossovers in accordance with its own algorithm.
As a general rule, crossovers can always be set higher than where they are set automatically by the Audyssey calibration. This is a user-preference issue, and may depend on what sounds (or measures) best to a particular individual. It is not a good idea to set crossovers lower than where they were set automatically during calibration, because those speakers will not be receiving any benefit from room EQ, somewhere a little below the original crossover point. They will also be down -3db in measured SPL at the point where the EQ stops. And, they will continue to play softer and softer as the frequencies go lower, so, they will not be providing much audible benefit at that point, anyway.
In addition, running speakers with crossovers below the original calibration setting, may consume valuable amplifier power, and may result in some distortion, while the more powerful subwoofers are being correspondingly under-utilized. So again, it is typically better not to reduce crossovers from wherever our AVR's set them. Crossovers of at least 80Hz are typically recommended in THX standards, and for general best-practice purposes, as subs will nearly always do a better job of reproducing the mid-bass frequencies up to 80Hz or so. 80Hz is also used as a standard frequency for where many people will not be able to localize a subwoofer.
*** Sometimes, HT owners may be a little reluctant to set crossovers of 80Hz, or higher, due to concern that they won't be using the full capabilities of their tower, or large bookshelf, speakers. But, in considering that, it is important to understand how the crossovers in our AVR's actually work. Crossovers are not like brick walls, where the speaker suddenly stops playing everything below 80Hz, and the subwoofer suddenly starts. We might be able to hear that kind of abrupt transition from speaker(s) to subwoofer(s).
Instead, when we set a crossover, the AVR implements a high pass filter (HPF) for the speaker(s) and a low pass filter (LPF) for the subwoofer(s). The high pass filter is designed to pass all of the frequencies above that point, while the low pass filter passes all of the frequencies below that point. But, both filters have a slope which gradually reduces the volume of the speaker or subwoofer to insure a smooth transition at the crossover point. (The filters also gradually reduce the volume to insure that neither speakers nor subwoofers are damaged by trying to play frequencies that they shouldn't be playing, at anything approaching full volume.)
The HPF for the speaker(s) is typically a 2nd order filter with a slope of -12dB per octave. The LPF for the sub(s) is typically a 4th order filter with a slope of -24dB per octave. In theory, at an 80Hz crossover, the speaker would be playing at -3dB, and the subwoofer would be playing at +3dB. That would maintain equilibrium at that frequency, so that there would be no apparent change in volume there, while allowing both transducers to start their roll-offs on the other side of the crossover.
(As is noted elsewhere in the Guide, at somewhere near the crossover is also where there can sometimes be phase-cancellation between speakers and subwoofers since they are playing the same content at about the same volume level. where that occurs however, the area of cancellation is often quite small, and there is no audible effect. It should also be noted that strong subwoofer boosts can raise the area where phase-cancellation might occur, since the subwoofer is playing much louder than the speaker at 80Hz, for instance, but closer to the same volume at 90 or 100Hz. This is not an issue that we should normally be concerned about unless we hear something that we shouldn't. Once again, our personal hearing preferences should be our final judge of which crossovers to use.)
Putting the previous example of how crossovers work into practical terms, with an 80Hz crossover, a speaker would still be playing 60Hz content, but a little softer than it plays 80Hz content. It would play 60Hz at about -8dB, and 40Hz content about -12dB softer than it was playing at just a little above 80Hz. The frequencies between 40Hz and 80Hz would constitute one octave (40Hz times 2).
[Note: An octave, in this context is a series of 8 notes, where the frequency of one is twice that of the other. So, from 40Hz to 80Hz would be an octave. From 80Hz to 160Hz would be one octave.]
Looking at that same 80Hz crossover, from the standpoint of the subwoofer, the sub would still play 100Hz and higher frequencies, but at gradually decreasing volumes. By the time the it reached 160Hz, it would play -24dB softer. The subwoofer's -24dB octave would consist of the frequencies between 80Hz and 160Hz (80Hz times 2). Once again, strong subwoofer boosts would slightly affect the attenuation of bass above the crossover. In general, the crossover filters, set by the AVR, allow the speaker to gradually give way to the subwoofer, while allowing both subwoofer and speaker to play a little above/below the selected crossover point. It's all about trying to achieve a more blended transition.
The crossovers within our speakers are similarly designed to create gradual transitions from woofer to mid-range, and from mid-range to tweeter (in a three-way speaker), although the precise amount of attenuation employed in the filters may vary among different speaker designs. When employed properly, crossovers enable us to listen to our audio content with no audible transitions at all from driver-to-driver within a speaker, or from speaker-to-subwoofer within an audio system.
As far as setting crossovers within our audio systems is concerned, there may be circumstances in which a crossover even lower than 80Hz is desirable, where measurements, or our own perceptions of sound quality, guide us. A situation where our speakers were set to Large during the calibration (because the measured F3 point was in the mid-thirties or lower) might lend itself to setting a crossover lower than 80Hz. For instance, in some cases, it is possible that a 60Hz crossover might provide an apparently smoother transition, or some other audible advantage, compared to the standard 80Hz crossover.
But, concern that we are wasting the capabilities of our speakers should not be a strong factor in our decision regarding what crossovers to use, as the speakers will always be playing content somewhat below the crossovers we select. It is also helpful to remember that our speakers have to play a very large frequency range, and that relieving them of some low-frequency burden (which requires a disproportionate amount of the amplifier power assigned to that channel) may help them to play their entire frequency range more effectively, and with less potential distortion.
Our subwoofers, on the other hand, have just that one specialized job to perform--playing low frequencies. And, as noted earlier, they typically do that single specialized job much better than even very large tower speakers. Particularly where bass-heavy movies or bass-enhanced music are concerned, crossovers of about 80Hz or higher are usually likely to improve our sound quality. In considering the use of crossovers for 5.1 movies, and for bass-enhanced music, it is important to realize how much additional low-frequency demand may be placed on our speakers. That is especially the case at higher master volume levels.
It is also worth noting that we may be better able to mitigate negative bass influences with our subwoofers than we can with our speakers. We will usually position our front speakers in specific locations on the front wall where they form an equilateral triangle with our listening position, or where they look good aesthetically. As we drop below 120Hz, and especially below 80Hz, room modes may be affecting more and more of our bass response. In some cases, we may be able to position our subwoofer(s) more strategically to sound good just for bass frequencies, where our front speakers have to be positioned to sound good for all frequencies.
If the subwoofers are better positioned with respect to room modes than the front speakers, then transferring more of the bass load (below about 80Hz) to our subwoofers may help to promote a superior frequency response with better sounding bass. This is always an issue that has to be resolved on a room-by-room basis, but it helps to understand some of the reasons behind the typical advice to use 80Hz or higher crossovers.
I think that the more that we understand how much more capable our real subwoofers are compared to the woofers in our speakers, the easier it is to use slightly higher crossovers. It isn't just a matter of woofer diameter, it is also a function of the subwoofer driver's motor strength and excursion capabilities (literally moving in and out to displace air and create SPL), the cabinet volume, the amplifier power, the DSP employed, and even the tuning point of a ported speaker, compared to both sealed and ported subwoofers. The subwoofer is typically far better at playing frequencies below about 80Hz or so, with higher SPL and less distortion.
A final crossover issue which is worth exploring involves bass localization. A subwoofer can be localized when the bass sounds that we hear are obviously coming from the subwoofer itself. Early listening tests seemed to indicate that most people couldn't locate the specific direction of bass sounds below about 80Hz (or a little higher in most cases). At very low-frequencies, bass sounds are definitely not as directional as they are for frequencies in our more normal listening range.
It was for that reason that 80Hz was originally chosen as a recommended crossover point from speakers to subwoofers. This is a fairly complex subject, in its own right. In general, where someone cannot identify bass sounds as coming directly from a subwoofer, an 80Hz or slightly higher crossover may be a very good choice. This issue of bass localization is explored in more detail in Section III-D.
Section III-B: Low Frequency Effects Channel:
There is another setting, associated with our subwoofers, which it is important to mention in the context of this discussion. It is not a crossover, but it does control the content we hear in the LFE channel. The .1 LFE (low-frequency effects) channel is a separate bass channel, played only by subwoofers, as long as there are any subwoofers configured in an audio system. (If there are no subwoofers connected to the system, and turned on in the configuration menu of an AVR, then the front speakers will automatically be set to Large, and LFE content will be routed to those speakers.)
The LFE channel is intended to give audio mixers an opportunity to add more bass SPL to 5.1 audio tracks. As explained in several sections of the Guide, we don't hear bass frequencies as well as we do other frequencies in our normal hearing range, so adding even more low-bass SPL, via a separate channel, can really enhance bass sound effects in movie soundtracks.
[It may be worthwhile to distinguish between two-channel music--stereo-- that has been bass-enhanced electronically, and 5.1 channel music. Some bass-enhanced music can contain very low-bass frequencies, and even bass sine waves, and it can be recorded at higher than normal volumes. The subwoofers will handle the low-frequency content in that music via the normal crossovers. But, only 5.1 music (or of course movies) will have an LFE channel which is dedicated to the subwoofers, and which is programmed to be 10db louder than the regular channels. Two-channel music which has been up-converted to surround sound, via Dolby Pro Logic, or by some other surround mode, is still two-channel music and doesn't have an LFE channel.]
As noted in other sections, the LFE channel has a max SPL of 115db for peaks, compared to 105db for the regular channels. That provides an additional 10db of bass SPL for audio mixers to be able to add to their 5.1 tracks. And, that allows selected low-bass sounds to stand-out more. In addition to the obvious difference of 10db between the regular channels and the LFE channel, however, there are also some differences in bass content.
The bass content in the regular channels ranges from the upper-end of the bass spectrum (500Hz) to the lowest frequencies contained in a recording. So, the range could be from 500Hz down to as low as single digits (~2Hz to 5Hz). The great majority of that bass content would be played by the woofers in the regular channels, with the subwoofers taking over at the crossover from the speakers to the subwoofer(s).
The LFE channel has a much more restricted and concentrated frequency range, and is exclusive to the subwoofers, if there are any in an audio system. Bass content in the LFE channel may still go down to the low single digits, depending on the film, but there is a filter which attenuates the bass content above about 120Hz. But, even beyond the difference established by the restricted frequency range, most real content in the LFE channel appears to be concentrated below about 80Hz. And, I believe that is generally intentional.
It is important to understand that the LFE channel is specifically designed to give more weight to the low-bass (and not quite as much to the mid and higher bass) in scenes where low-frequency effects are deemed appropriate by the sound mixer. So, the LFE channel may not be in very obvious or continuous operation during many movies. In some cases, the LFE channel may be used sparingly, and it may just kick-in with more really low and loud bass, whenever a sound mixer wants to emphasize the sound effect at a particular point in a film. In other movies, with sustained and intense low-bass, the LFE channel may be extremely involved throughout the movie. I believe that is the case in Batman Versus Superman, for instance.
* The LFE channel, which exists only in 5.1 (or higher) movies and music, has it's own setting in our AVR's, called the LPF of LFE. Since the LFE channel is intended to contain bass content up to 120Hz, the typical setting for the LPF is 120Hz. And, that is the customary setting which most AVR's engage by default. However, a number of audio experts, including Mark Seaton and Roger Dressler (formerly with Dolby Labs and one of the creators of Pro logic II) believe that it can make sense to experiment with lower LPF settings.
Some people have suggested that relatively little meaningful bass content is mixed into the LFE channel above about 80Hz, as the LFE channel is primarily intended to emphasize lower bass sounds and special effects. That may or may not be correct in general, although some film mixers have indicated that most of their meaningful LFE content is in the lower bass range. It would make sense for that to be the case, since the original Dolby/THX standard was for speakers to crossover to subwoofers at 80Hz, and the <80Hz frequencies were always the ones most commonly associated with subwoofers.
In some cases, setting a lower LPF might emphasize low-bass frequencies a little more, and might also result in slightly clearer bass. Since the LPF is simply a filter, which gradually attenuates volume levels, setting a lower LPF will not completely eliminate bass above the filter, but it will roll-off the higher bass content a little earlier. For instance, an LPF setting of 80Hz would roll-off the 100Hz frequencies by 6dB, and the 120Hz frequencies by 12dB. Doing that would provide relatively more emphasis to the low-bass frequencies, compared to the mid-bass frequencies. That is similar to, but probably more subtle than, approaching the bass from the bottom by lifting the lowest frequencies with a rising house curve.
Mark Seaton has made the point that the more someone is boosting his subwoofer(s), the more that an 80Hz LPF may be helpful in making the bass blend well with the speakers in an audio system. That would particularly be the case where someone was using 80Hz crossovers for the regular channels. Remember that a subwoofer boost lifts all of the bass frequencies symmetrically, in both the regular channels and in the LFE channel. Where significant subwoofer boosts are employed, the bass frequencies above 80Hz in the LFE channel (which are already 10dB louder than the regular channels) might seem to stand out too much in comparison to the lower bass frequencies. Again, that might be more likely to be noticeable where 80Hz crossovers are employed for the regular channels.
Some people may notice a little more bass clarity, and a little greater concentration on the low-bass, with an 80Hz setting. Others may prefer the fuller mid-bass sound with the default 120Hz setting, or may perhaps prefer a compromise setting of 90Hz or 100Hz. The differences among the various settings are probably fairly subtle, depending on the listener, and which setting sounds better is strictly a user preference issue. Although AVR makers typically employ a default LPF setting of 120Hz, there is no absolute right or wrong way to use the LPF of LFE. (FWIW, I do think it is possible that a higher LPF setting might contribute to subwoofer localization, when bass-heavy 5.1 content is playing.)
As a general rule, there may be no particular reason to experiment with the LPF unless a fairly significant independent sub boost is employed--perhaps at least 3 or 4db on top of DEQ, or even more than that without DEQ, and unless crossovers of 80Hz or 90Hz are employed for the regular channels. Or, unless someone is specifically looking for greater clarity in the mid-bass range. There is some additional discussion of methods to achieve mid-bass clarity in the next section on Cascading Crossovers.
[Some additional discussion of the LPF of LFE, including comments from Mark Seaton and Roger Dressler, can be found in the Audyssey FAQ, linked below. It should be noted, however, that one summary comment by another AVS member, at the very end of that discussion, is not correct. LFE material is not "brick wall" filtered at 120Hz. As noted above, the LPF (at any setting) simply rolls-off content gradually, just as any other low-pass filter does.]
"Official" Audyssey thread (FAQ in post #51779)
[It should also be noted, that just as some AVR's only offer crossovers which are already fixed at 80Hz, some AVR's do not allow LFE adjustments. Yamaha AVR's, for instance, do not have variable LPF of LFE settings. In that case, the default setting will be 120Hz.]
Section III-C: Cascading Crossovers:
The concept of using cascading crossovers to increase mid-bass clarity, and to increase dialogue intelligibility, is one that has been around for a while. It may be especially helpful where someone is using significant subwoofer boosts in order to emphasize low-bass frequencies, or to emphasize mid-bass chest punch.
The process is typically defined as setting two crossovers in different places, such as in your AVR and in your subwoofer, to combine at the same frequency. In this case, we won't actually be setting a "crossover" in the subwoofer, although it is often labelled as that on the subwoofer amplifier. We will just be setting a low-pass filter in the subwoofer which corresponds to the crossovers in our AVR. (As explained earlier in Section III, a low-pass filter "passes" frequencies below the set point. By doing that, it regulates the frequencies which a subwoofer is allowed to play.)
As with all setting options, cascading crossovers is something which is implemented after an audio system is calibrated. So, the settings described below are changed from the default settings after the auto-calibration routine is performed.
There are three components to cascading crossovers. First, there are the crossovers from the speakers to the subwoofers, which are typically set at about 80Hz. Surround and height channels may have higher crossovers than that, and that is usually fine. It is mainly the front soundstage, which carries most of the meaningful content and all of the dialogue, that we are trying to affect.
Second, there is the LPF of LFE, which controls a separate bass channel (the .1 low-frequency effects channel) as explained in the previous subsection. Only the subwoofers play the LFE content, and that additional bass content is only present with 5.1 movies and 5.1 music. To implement cascading crossovers, both bass sources in the AVR would be set to the same ~80Hz frequency. So, the LPF of LFE in the AVR would also be changed to 80Hz.
(Some AVR's, such as Yamaha AVR's, don't allow the LPF of LFE to be changed. It always remains at the default setting of 120Hz. If so, it is no problem. Setting the LPF in the subwoofer itself to 80Hz will still have full effect on the crossovers from the speakers to the subwoofer, and the low-frequency effects channel will still roll-off a little faster, too. So, the concept of cascading crossovers will still work.)
The third component is the low-pass filter (LPF) in the subwoofers themselves. As noted above, that filter may be labelled as a "crossover" on the subwoofer's plate amp. It controls how high in frequency the subwoofer is allowed to play before starting to roll-off. To make the two bass sources in the AVR cascade, it would also be necessary to set the subwoofer(s) low-pass filter to the same ~80Hz frequency.
That will often be done with an analogue knob on the plate amp. There may be an "On/Off" switch or an "In/Out" switch which allows users to engage their own LPF. If there is such a switch, setting it to "On" or "In" depending on the switch, will enable the analogue knob to control the sub's low-pass filter. Cascading crossovers occur when both bass sources in the AVR approximately correspond to the LPF ("crossover") in the subwoofer.
(If the analogue knob on a subwoofer doesn't allow exact adjustment to 80Hz, just try to get close. Exact correspondence doesn't matter. If the subwoofer started rolling-off faster at 85 or 90Hz, instead of right at 80Hz, the audible result would still be virtually identical.)
The practical effect of using cascading crossovers is to cause the subwoofers to roll-off faster above the selected crossover point. As explained in Subsections III-A and III-B, the crossover which tells the subs where to take over from the speakers is not a brick wall. It contains a low-pass filter which causes the subwoofer to roll-off gently above a specific frequency, such as 80Hz. That roll-off is typically -24db per octave. When we implement cascading crossovers, we are making the subwoofers roll-off faster, above a certain frequency, so that less bass will leak into frequencies above that crossover. As explained later in this section, that bass leakage into frequencies above about 80Hz can sometimes affect male voice clarity, among other things.
* I decided to add a little more detail to this idea of rolling-off subs a little faster, above 80Hz, especially where significant subwoofer boosts are employed. Let's look at the LFE channel first and assume an LPF setting of 120Hz. For the 8-note octave between 120Hz and 240Hz, the low-pass filter will roll-off the subs by 24dB. That sounds like a lot, but what that means in practical terms is that the subwoofer will gradually lose about -21dB between 120Hz and 240Hz.
Now, let's assume that someone wants to use a fairly significant subwoofer boost. An 8dB boost using something such as DEQ, or through independent subwoofer boosts, would not be at all uncommon. Instead of being down by 8 or 9dB at 150Hz, the subwoofer would still be playing that frequency at about the same volume level that it would have been playing, if there hadn't been a boost. And, the LFE channel is already playing 10dB hotter than the regular channels.
It is easy to see that the subwoofer boost, occurring above 120Hz could make the bass in the LFE channel sound a little heavy. It is also easy to see how that boost above 120Hz could make the subwoofers strain a little more, depending on the overall listening/subwoofer volume. Very few subwoofers can play as clearly, with as little strain or compression, at 160Hz as they can at 120Hz. Rolling-off the LFE channel, above 80Hz, can help to alleviate that potential issue.
The same thing happens in the regular channels, but starting at a lower frequency, if an 80Hz crossover is being employed. Just a 6dB boost (which is very modest for some people) would make the 100Hz frequency play about as loudly as the 80Hz frequency would have played without the boost. Again, it is easy to imagine that the extra boost for the center channel (above 80Hz) could make some male voices sound a little more chesty, as bass fundamentals were amplified. And, that in-turn, could make dialogue sound a little thicker and a little less intelligible. Again, rolling-off the subs a little faster, above the crossover, helps to alleviate that issue where it is an audible problem.
In many cases, listeners have found that rolling-off the subwoofers faster improves overall mid-bass clarity, and especially dialogue clarity. It may also, in some cases, concentrate the bass a little more strongly below the crossover. In my opinion, cascading crossovers are most likely to work well, where the three speakers on the front soundstage are reasonably capable of handling frequencies above about 80 or 90Hz.
I personally believe that unless the speakers on the front soundstage can play 80Hz or so with reasonable power and low distortion, we may be better off setting higher crossovers and letting our subs play frequencies higher than 80Hz. And, we also may not want the subs rolling-off any faster at 80Hz. For that reason, very small bookshelf speakers might not be good candidates for cascading crossovers.
[Combining two crossovers may potentially cause some cancellation at that specific frequency (such as at 80Hz) in some cases, although that may or may not be audible if it does happen. FWIW, I believe that cascading crossovers, acting on their own, are not very likely to cause cancellation at the crossover, or to create an audible problem even if a narrow range of cancellation does occur.
However, as noted in an example on the Guide thread, by @bscool, if some measurable cancellation does occur, it can typically be corrected by either adjusting phase or subwoofer distance, as illustrated at the end of this subsection. In any event, cascading crossovers will typically increase the strength and clarity of the mid-bass frequencies as a whole. And, many other listeners who have tried the process have reported an overall improvement in sound quality and in mid-bass impact.]
Some time ago, I decided to experiment with the concept of cascading crossovers in my system, and I liked the results very much. I will explain what I did and what I liked about the way it influenced my sound quality. (I will note at the outset, that for anyone using Audyssey or some other form of room EQ, nothing about implementing cascading crossovers interferes with the filters set by automated room EQ. This is strictly a post-calibration tweak.)
First, I should explain that I have very capable speakers in my 7.1 system, and I never use crossovers higher than 80Hz. (I think that having reasonably capable speakers, which can handle crossovers below about 100Hz may be a prerequisite for successfully implementing cascading crossovers.) Second, I also prefer to use an LPF of LFE setting of 80Hz. I get better bass clarity when I set my LPF to 80Hz, rather than to the default 120Hz. Potential advantages to using the lower LPF are briefly described at the end of the previous subsection, and in greater detail in the Audyssey FAQ. The fact that I was already using a lower than typical LPF of LFE made me think that I might be a good candidate to try cascading crossovers.
Since I was already getting a very smooth transition at 80Hz, from my speakers to my subwoofers, I decided not to set my subwoofers' internal low-pass filters to that same 80Hz. Instead, I chose 100Hz, for my initial experiments, and then later tried 90Hz and 80Hz. Before attempting to explain what I experienced when I tried this, I should explain the physical mechanism involved. As discussed earlier, when we set a crossover for our speakers, in our AVR's, the speakers typically roll-off below the crossover at 12dB per octave, and the subwoofers roll-off above the crossover at 24dB per octave.
As noted earlier, in theory, the speakers will already be playing -3dB at 80Hz, and the subwoofers will be playing +3dB. But, the subwoofers are still playing the content above 120Hz, although at a reduced volume level, and their SPL still contributes to the overall sound and consumes some headroom from the subs. When, I set a 24db per octave, 100Hz LPF in my subwoofers themselves, I didn't affect the frequencies below 100Hz. But, I increased the magnitude of the roll-off occurring above 100Hz.
What I found when I tried this was that my mid-bass frequencies (up to 100Hz) seemed relatively louder than they had been, and my overall bass clarity improved. I especially noticed that I didn't have to boost my center channel as much as I had been doing, in order to hear clear dialogue. I think this is due to two factors. First, the higher bass content that had been played by my subwoofers was making the front speakers and surrounds a little heavy-sounding in proportion to the somewhat smaller center channel. And, second, since I was already using a heavy subwoofer boost, cutting-off the subs a little quicker imparted less bass coloration to the voices coming from the CC.
This is one of the reasons that I personally prefer not to use DEQ. I don't like boosting the bass in the center channel, with the voice coloration that I notice when I do that. Deep male voices typically only go down to a fundamental frequency of about 90Hz, so bass boosts above that frequency may make men's voices sound unnaturally thick and chesty to some people. As noted in other sections, however, whether we notice that sort of thing, or care about it, is strictly a YMMV issue. (I make up for not using DEQ by implementing a much more substantial subwoofer boost for movies.)
** I also decided to add a little more detail to the explanation of why we may hear more mid-bass and overall clarity when bass boosts don't go above about 80 or 90Hz. Using voices is an excellent way to describe what I think is happening, and that is where I personally notice the additional clarity the most. The human voice is an instrument with a large frequency range. I said that bass boosts above 80 or 90Hz may potentially make male voices sound "chesty". In vocal music, a chest sound is deeper and more resonant than a head tone, which is produced higher in the voice box. The chest tone requires more air, and it resonates lower in the voice box than the head tone does, but it can also sound "throatier", and it has less clarity or "brilliance".
Some consonants, such as "B", "C", "D", "G", "T", "V", and "Z" which all share the same long "ee" sound, may be more difficult to distinguish if they are pronounced with too much chest tone. Some vowels can also be harder to distinguish if more bass sound is added to them, because the voice will sound slightly thicker. I believe that is especially the case if the person speaking has a strong accent, or if he fails to articulate clearly, or if ambient noises in the soundtrack make voices harder to hear clearly to start with.
(When someone articulates, he says each syllable of a word clearly and distinctly. James Earl Jones is a great example of a person with a very deep and resonant voice who is nevertheless very easy to understand. But, he had a speech impediment as a child and worked very hard to learn to speak slowly and with excellent articulation. Most actors do not have that style of speech and that kind of articulate diction.)
Remember also that if subwoofers are strongly boosted, with the normal 80Hz crossover in the AVR, they are only rolling-off at 24db per octave above 80Hz. So, at 100Hz, the subwoofer has only rolled-off by 6db and can still provide quite a lot of bass coloration to male voices. To me, that can make the voices sound a little unnatural as well as more difficult to understand. So, where I may not mind a little additional bass resonance in some music (the cello or the kettle drum, for instance), I may not like it quite as much for some other instruments. And, where I absolutely want it for the low-bass special effects in movies (well below an 80Hz crossover), I may not want that extra resonance at all where the human voice is concerned.
I found that as I implemented cascading crossovers at 100Hz, and then at 90Hz, and finally at 80Hz, I was able to concentrate a little more bass below 100Hz, and then below 90Hz, and then below 80Hz. And, not only did the mid-bass clarity improve with each attempt, but my mid-bass tactile response also increased as a result. That chest punch sensation is explained in detail in Section VII, but briefly, most people seem to feel the sensation most strongly between about 50Hz and 100Hz.
There is some evidence that the sensation may peak for most of us at around 63Hz. That specific number was the conclusion of one study I read years ago, and some subwoofer makers, such as SVS, provide the capability to add a pre-programmed peak at that frequency into their higher-end subwoofer models which have advanced PEQ. If we make our subwoofers roll-off more quickly above 80Hz, by implementing a 48dB per octave filter, we are doubling the roll-off.
So, although there is still some transition between speakers and subwoofers, the subwoofers have rolled-off a good deal more at 100Hz, and they have rolled-off by about an extra 12dB at 120Hz. It is easy to understand how larger subwoofer boosts would allow us to benefit from a quicker roll-off above our selected crossover. And, it is easy to understand how we might be increasing the punchiness of the bass in the range where most people feel those chest punch sensations most strongly.
I offer this method of cascading crossovers as a means of potentially obtaining additional mid-bass SPL and chest punch, combined with potential improvements in overall bass clarity. (The clarity was the real key for me, but again, I use a lot of subwoofer boost for movies.) Determining where to set the LPF in the subwoofers themselves, and what slope to select if that is an option, is something which may require some individual experimentation. But, in my personal opinion, it may turn-out to be an excellent solution for someone wanting to maximize mid-bass SPL and clarity. The setting procedure is summarized as follows:
Setting Procedure:
To recap the procedure to follow in setting cascading crossovers, the following three steps would be performed after a calibration:
First, all of the speaker crossovers in the front soundstage would need to be set not higher than about 100Hz, in the AVR, and 80Hz or 90Hz would be better. So, we would need to have fairly capable speakers for at least our three speakers on the front soundstage. (It might not matter quite as much for surrounds, rear surrounds, height speakers, and so on, as it would be for the channels which carry so much of the fundamental content of both movies and music.)
Second, we would ideally need to be able to implement an LPF of LFE which approximately matches our speaker crossovers. Let's just say we are using 80Hz to make things simple. (If we couldn't adjust the LPF--as is the case in some Yamaha AVR's--we could still try cascading crossovers anyway, as explained earlier.)
Third, we would need to be able to implement a low-pass filter in the subwoofers themselves, or with a miniDSP, which would approximately match the crossovers to our speakers. (Most subwoofers will have either digital DSP, or an analogue knob--sometimes labelled Crossover or LFE--which will enable us to set a low-pass filter for the sub. Some subwoofers may also enable us to manage the slope of that filter. In my case, I rolled-off the bass above 80Hz at 24dB per octave.)
When we implement all three of those settings to coincide: the speaker crossovers, and the LPF of LFE in the AVR; and the LPF in our subwoofer(s), the subwoofers will roll-off much faster above our target frequency of let's say 80Hz, and the mid-bass SPL and tactile sensations will be more concentrated below that frequency. I think it would be generally preferable to make the three low-pass filters correspond with each other, but they don't have to correspond exactly (especially with an analogue knob). And, if the LPF of LFE in the AVR can't be adjusted, there is still some benefit to making the speakers' crossovers, and the low-pass filter in the subwoofer(s), correspond.
*** After trying the 100Hz LPF in my subs for a few days, I experimented with dropping the LPF in the subwoofers from 100Hz to 90Hz, and the results were even better. In my particular case, the bass frequencies were even more distinct, and the center channel was even clearer. In fact, I was able to reduce the volume on my CC by about 2dB, depending on the movie, and still understand dialogue perfectly well. I like using a large subwoofer boost for the very low-frequencies. I am in a large room on concrete, and it can take a significant subwoofer boost to generate the low-frequency sounds and tactile sensations I like.
But, using a large subwoofer boost also tends to make voices slightly thicker, and a little harder to understand, as explained above. We may get so accustomed to a slight bass coloration in voices, and in other mid-bass sounds, that we may not even notice that we are hearing it any more. At least I didn't. But, when that bass coloration is removed, the resulting sound from all of the speakers is much clearer. And, I can especially hear the difference in the center channel.
After getting used to the 90Hz LPF for a few days, which is how I typically like to test setting changes, I decided to drop the LPF in my subwoofers to 80Hz. At that point, all of my settings aligned at 80Hz. With each change, going from full-range settings in the subs to 100Hz, and then successively down to 90Hz, and then to 80Hz, I liked the results better. The acid test for me was when I watched Battle Los Angeles with the all 80Hz settings. It had been a couple of years since I had seen it, but I remembered how difficult it was to hear some of the dialogue during battle sequences. The mid-bass and low-bass were more impactful than ever, and the dialogue was easily understandable at lower volume levels than it had been before, even when I had boosted the CC.
The cascading crossovers make a noticeable difference to me, and to several dozen others who have reported trying them since I first wrote this. I think that the more subwoofer boost we use, the more that this approach may be helpful for us. This is probably not going to be a good solution for everyone. Our audio systems, and individual preferences, are just too diverse for any single method, or setting, to be successful for everyone. But, I definitely recommend trying it, if someone believes that he may be a good candidate, and if his subwoofers allow him to set a lower LPF, which he believes might potentially correspond well with the crossovers he is using in his AVR.
(Just to reiterate, none of the changes I am suggesting will interfere with the automated room EQ that you may be employing, and any of these changes are implemented after a calibration. You would always want room EQ to measure the full capabilities of your speakers and subwoofers. It is only after room EQ has set filters for the various channels that you would implement any limitations to the upper range of the subwoofers.)
Dealing with Cancellation:
I decided to put a brief description of the cancellation that people may sometimes encounter when they incorporate subwoofers into an audio system. Due to a variety of factors, some frequencies may cancel each other, eliminating bass at those frequencies. Cancellation can occur with a single subwoofer interacting with a room, or with dual subwoofers interacting with each other, or with subwoofers and speakers interacting at a crossover. That cancellation may occur either with or without cascading crossovers.
(A separate discussion of dealing with phase cancellation, between the subwoofers themselves, is offered in Section IV-B.)
Cancellation always looks bad on a graphed frequency response. (It looks like a a deep V shape in an FR graph.) And, in general, we would prefer not to have any cancellation at any frequency. But, some cancellation may be pretty inevitable in most HT rooms, even where measuring capability and methods of independent EQ are available. Typically, where we have those capabilities, we are trying to move cancellation to another part of the frequency range, where it is less audible, or to another part of the room, where no one is sitting. Perfect bass in every part of a room is extremely difficult to achieve, even where we have four optimally-situated subwoofers.
I think it is worth pointing-out, for people who may be reading this and who do not have measuring capability, that not all cancellation is even audible. That is particularly the case with relatively narrow areas where frequencies are cancelling. For example, let's say that we have some cancellation centered at about 80Hz. That would be a fairly common area to expect some cancellation from speakers interacting (playing the same frequencies) with subwoofers.
We might not be able to hear that cancellation at all unless it covered a wide area. For instance, the octave between 60Hz and 120Hz consists of 8 notes, so each note in that octave covers slightly more than 10Hz. Most sounds we listen to are very complex, consisting of multiple notes (or frequencies, if the sounds aren't musical in nature), and harmonics (overtones) of those sounds. And, our brains are very adept at filling-in missing information. So, a narrow area of cancellation, centered on 80Hz, might not be audible at all.
Wide areas of cancellation, spanning more than 10Hz, might be more audible, as reduced bass in that particular area of the frequency response. But, for someone who doesn't have measuring capabilities, this is not necessarily something to be concerned about with our audio systems (either with, or without cascading crossovers) unless we suspect (either by listening or measuring) that we are losing some significant bass somewhere.
If we do have some reason to believe that we are experiencing cancellation, either via measurements, or from hearing a specific area of the FR where we think there is less bass SPL than there should be, there are a couple of different methods to reduce or eliminate the cancellation occurring at that frequency. One way would be to adjust the phase control on one or more of our subwoofers. Another way would be to adjust the subwoofer distance control in our AVR's. In either case, we could try playing an 80Hz test tone, or using some steady bass content at about that frequency.
Whether to use phase or distance settings probably depends on what kind of subs someone has (and how they are configured); and on what kind of AVR he has. If a single sub has a phase control, adjusting the phase on that subwoofer may be the easiest way to remove cancellation at a crossover. If a subwoofer doesn't have a phase control, adjusting the subwoofer's distance setting in the AVR will work. If someone has two equidistant subs, Y-connected into a single sub out in an AVR, then the use of the AVR's distance control might be easier to use.
If we are using the phase control on a subwoofer, to make it integrate better with speakers, or with another subwoofer for instance, the following information may be helpful. As I understand it, a distance change of 1/2 wavelength corresponds to a 180 degree change in polarity. So for instance, for cancellation occurring at 80Hz (which is a wavelength 14' long), a distance change of about 7' should reverse the phase completely. If a different crossover is used, there are online calculators which make it easy to correlate frequencies with wavelengths.
The real key to remediating cancellation, though, is to adjust the phase or the distance control gradually, measuring as you go, to determine what setting makes the cancellation either disappear (perhaps by moving it to a more remote area of the room), or move to a higher (or lower) frequency where it will have less audible effect. In an extreme case, and where the ability to measure the FR is not available, it might be possible to hear the bass getting stronger with different phase or distance adjustments, while playing some steady bass content.
The second method, the sub distance tweak, is offered compliments of @Alan P. Both the phase change, and the distance tweak, ideally require the use of measuring equipment such as REW. In the absence of more sophisticated measurement abilities, the use of an SPL meter would be helpful. You would just be measuring a relative increase in the volume level, at the crossover, with changes in either phase or distance settings.
It would be more difficult to do the distance tweak, or the phase adjustment just by listening, unless cancellation at the crossover were quite audible, and that would be rare. (Of course, if it's not at all audible to start with, then many people may not want to adjust the phase, or to perform the sub distance tweak, at all.)
Sub Distance Tweak:
1. Measure the center channel and the subs with REW or comparable software (REW HDMI CH3) using a test tone of 80Hz, or whatever corresponds to the crossover.
2. Add to the sub distance setting of one subwoofer (or of both subs equally, if using an AVR which has a single subwoofer distance setting) in 1' increments. (With some AVR's, you must make sure to back-out of the distance setting menu before the new setting will take effect.)
3. Remeasure.
4. Repeat until you get the smoothest transition around the crossover.
5. If using an 80Hz crossover, it would not typically be beneficial to add or subtract more than about 7' of distance. That corresponds to one-half of an 80Hz wavelength, and a change of one-half wavelength would change the phase of a subwoofer by 180 degrees.
6. We normally have to choose between a good transition for the center channel and the subs, or for the front speakers and the subs. If someone is primarily interested in movies, balance the compromise in favor of CC+sub; and for music, measure with the L/R+sub.
It should be noted again that the procedure outlined above may not be necessary unless someone is either sufficiently curious to measure his results, and discovers something specific in the frequency response, or unless someone hears something that leads him to suspect that significant cancellation could be occurring.
Section III-D: Bass Localization:
As noted in previous sections, 80Hz was chosen as a standard crossover, for Dolby/THX purposes, because it was believed that most people wouldn't be able to distinguish bass sounds as specifically coming from a subwoofer, at frequencies around 80Hz, or just a little higher than that. But, was that assumption really correct? Can most of us really not distinguish directionality in sounds below about 80Hz?
I believe that the often repeated statement that we can't hear directionality in bass sounds below about 80Hz is actually not correct. But, in defense of whatever original research was done, and of whatever conclusions were reached, I believe it was always assumed that there would be at least one subwoofer on the front wall, and that people would have difficulty in distinguishing between bass coming from a subwoofer on that front wall, and the speakers on that front wall.
Based on what I have read about how the concept of "Reference" was developed (as the product of consensus) I am not certain that there were any actual listening tests involved in the idea that listeners would not be able to localize bass below about 80Hz. I recall reading that one of the audio experts involved said that there wouldn't be any localization below about 100Hz, and 80Hz was selected to provide an additional safety margin.
But, if there were any actual listening tests involved, then I don't believe that they could have been very comprehensive, because putting a subwoofer on a side wall, instead of on the front wall, would probably have been sufficient to demonstrate that bass frequencies below 80Hz can be localized. The further that the subwoofer were from the speakers, in a larger room, the easier that the bass localization would be.
Over time, I suspect that the original assumption of a subwoofer on the front wall morphed into a generalized belief (or audio myth) that we can't localize bass at all under about 80Hz, or a little higher. I don't know for sure that early researchers made that assumption, of a subwoofer on the front wall, but it's a reasonable hypothesis to explain their conclusion. In any event, I believe that I can easily demonstrate that we can localize bass sounds under 80Hz, and I think that I can explain why, even if we couldn't, 80Hz crossovers would not be a foolproof solution to potential bass localization from a subwoofer.
Let's start with a music example, and to do that, let's talk about a jazz combo. I listen to a lot of jazz, and examples of what I am about to say abound. A typical jazz combo might involve a minimum of three instruments (counting a vocalist as a potential instrument) and would often consist of four or more. I will take a four-instrument combo as the usual one.
In a typical four-instrument combo, there would always be percussion (a drum set), and there would always be a double (or upright) bass. The drums would carry, and sometimes vary the rhythm, and the upright bass would also contribute to the rhythm and would provide a bass counterpoint to the music. Then, there would usually be a piano, and either a vocalist or saxophonist, or whatever. But, I want to concentrate for a moment on the kick drum, and on the upright bass.
Kick drums go down to a low-fundamental of about 50Hz, and an upright (or double) bass, with the standard four strings, can go down to about 41Hz. So, my first question is, when you are listening to jazz, can you tell where in the room the low sound of a kick drum is coming from? If you are using a surround mode especially, or if it is 5-channel content, can you tell that the drum set is in a particular location in your room, during that recording, and hear the kick drum coming from that location?
Can you hear the bassist strumming the strings on the upright bass, even when he goes really low? Frequently in jazz music, the bassist stays very low throughout most of the song. In some cases, that bass sound is just helping to carry the rhythm rather than trying to stand-out distinctively. But, if your eyes are closed, can you still hear it and point to where in the room the low-bass sounds are coming from?
The fourth string of an upright bass has a fundamental frequency of 41Hz. Even if the bassist doesn't make full contact with that string, the lowest note will be around 45 or 50Hz. Can you hear that the lowest bass sounds of the upright bass are coming from a particular spot in the room, and that the placement of that instrument doesn't change for the duration of the song? If you can, then I believe that simple fact demonstrates that we can indeed localize bass sounds below 80Hz.
This is actually a pretty easy observation for me, because I don't use my subwoofers for music listening at all. So, I am not dealing with any crossovers when I listen to music. In fact, I have six large (full-range), widely separated speakers, in a big room, that I use for music. All six speakers face me across about a 30' width, and a 130 degree arc to my front and sides. So, it's easy for me to pinpoint where specific instruments are coming from, when I am playing 5-channel content, and when I am playing stereo content with a surround mode such as PLIIx.
I had actually been enjoying jazz music that way for years before I started to put what I was hearing into the context of bass localization below 80Hz. And then, more time passed before I decided to write about what I believe is a pervasive audio myth--that we can't localize bass below 80Hz. I believe that any of us potentially can localize bass sounds below 80Hz, if our subwoofer doesn't happen to be on the same wall as our speakers. I think that would be even more likely in a larger room, where the speakers and the subwoofer were more widely separated.
I decided to add some non-music related examples to this discussion. Any of us could make a simple test of our ability to distinguish directionality in bass sounds by playing low-frequency test tones through a subwoofer, with the volume turned-down on our other speakers, or turned-up on our subwoofer. If we can hear test tones, at frequencies below 80Hz, distinctly coming from our subwoofer, then we are hearing directionality in bass below 80Hz.
In fact, when Audyssey or some other form of room correction plays the first series of test sweeps through each subwoofer, for purposes of level-matching the subs, can you distinctly hear those sweeps as coming from each individual subwoofer? I have always found that first series of bass sweeps (thumps) to be quite distinctive, as coming from each widely-separated subwoofer, in succession. The pink noise used in those first sweeps is the range from approximately 30Hz to 70Hz. If you hear that pink noise quite distinctly, as a single low-frequency thump coming from a specific subwoofer, then you are hearing directionality in bass below 80Hz.
Now saying that we potentially can notice something is not exactly the same thing as saying that we definitely will notice. And, that is true with bass localization too. After all, the frequencies played by our subwoofers are mixed together with the frequencies played by our speakers. And, most of the content that we listen to is complex content consisting of both fundamental frequencies and harmonics of those frequencies. Some of us may be more likely to notice than others, but complex content might still be a factor in whether we noticed any bass localization.
Our brains are also very good at adapting to familiar circumstances and expectations. That's why two widely separated speakers can create a phantom (stereo) image at the center point of the speakers. And, it's why we can position a center channel below the level of our screens, or large displays, and still hear the voices coming out of the mouths of the characters whose heads are near the top of the screens. We want to hear what we know we are supposed to hear, and what we are accustomed to hearing, and our adaptable brains do the rest.
But, once our subwoofers move away from where our speakers are located, the illusion that the bass sounds below about 80Hz are coming from the speakers becomes harder to maintain. That is why people who position a single sub on a side wall, or on the back wall, or even in one corner of the front wall, may often have trouble with 80Hz crossovers, or sometimes even with 60Hz crossovers. It's because the ability to hear directionality in bass sounds doesn't just magically go away below 80Hz.
As I demonstrated with the jazz example, or as anyone could demonstrate with a live orchestra or a live band performance, it's actually quite easy to localize bass sounds from percussion sections for instance, below 80Hz, and even below 60Hz. Another example would be a parade, where a band marches by and we hear the low thump of the bass drum as it approaches us on our left, passes us to our right, and recedes in the distance. Most of us probably just accept the myth that we can't localize bass sounds below 80Hz, and we don't apply our real world knowledge to the question.
Had I really thought about it much earlier, I would always have known from personal experience with live performances, that we could localize bass below 80Hz. But, even if there were some validity to the idea that we can't localize bass sounds below 80Hz, would an 80Hz crossover always work? I don't think it would. Let's first address this question by talking about subwoofer boosts. Remember that crossovers work by gradually reducing the volume of a sub, above the crossover, while reducing the volume of speakers, below the crossover.
But, what happens when we turn-up the volume of our subwoofers, relative to the volume of our speakers? Does the crossover still work in exactly the way it was intended, or does the subwoofer now play 100Hz and 120Hz content louder than was anticipated in the crossover design? And, if it plays 100Hz and 120Hz content louder than the crossover's roll-off intended it to, can we localize the bass even more easily in those higher frequencies than we could the 80Hz frequency? It makes sense that we would be able to.
I think that could be another reason that cascading crossovers are such a good solution for people wanting to use significant subwoofer boosts. Cascading the crossover helps the 80Hz crossover to operate more in the way it was intended to, in the case of significant subwoofer boosts.
Here's another reason that I think the 80Hz crossover may not entirely prevent bass localization, where a subwoofer is not located on the front wall, with the main speakers in our HT systems. I have always wondered how the LFE channel factors into this question of bass localization. Even if we accept the original premise that the 80Hz crossover, from speakers to subwoofer, prevents bass localization, what about the 100Hz and 120Hz bass in the LFE channel, which plays with 5.1 content.
That bass content is only played by our subwoofers, if our speakers are set to Small. And, the default setting for the LPF of LFE in most AVR's is 120Hz. To make LFE content potentially even more problematical with respect to bass localization, the LFE channel is already playing +10dB louder than the content in the regular channels, irrespective of any subwoofer boost. (Sub boosts would affect the regular channels and the LFE channel equally, so subwoofer boosts would make the LFE content stand-out even more.)
Couldn't we localize a subwoofer, based on the higher volume level content above 80Hz, from the LFE channel? Wouldn't the louder 100Hz and 120Hz content potentially contribute to bass localization, even if the 80Hz crossover did work somewhat for the regular channels? I have actually preferred using an 80Hz LPF of LFE for several years now, and my subwoofers are well-distributed around the room, so that question would be harder for me to test. But, it has always bothered me that we accept the idea that an 80Hz crossover prevents localization for the regular channels, while most of us are still using the default 120Hz LPF of LFE for the LFE channel, which is already playing bass +10dB louder than the regular channels.
I believe that there is one final factor to discuss with respect to potential bass localization, below 80Hz, and that is tactile sensations. I can't speak for anyone else, but for me, bass TR (tactile response) has directionality too. When I hear a sudden percussive bass sound, at the right frequency, and feel a thump in my chest, I am aware of the general direction from which the sensation originates. It may be helpful to remember that a significant study of chest punch determined that the average frequency where most people felt the maximum impact was 63Hz. And, at least two sub makers offer pre-programmed PEQ boosts at that specific frequency.
If we do feel directionality in bass tactile sensations well below 80Hz, then couldn't some of us localize subwoofers which were not on front walls for that reason too? I believe that we could. In fact, I believe that even very low-frequency tactile sensations--such as thudding and rumbling sounds/sensations can also have directionality. They certainly seem to outdoors, when something heavy falls to the ground and we feel the direction of the vibrations. They certainly seem to in my large room, as well, with my widely distributed subs. That is one reason I have four of them now. I didn't really need more than three subs for overall headroom or frequency response, but I always had a directional hole in the low-frequency bass and TR until I added the fourth subwoofer. I think that was partly due to the somewhat challenging geometry in my 6000^3 room, though.
I definitely think that some of us may be more sensitive than others to directionality in both bass sounds and sensations. But, I also think that there are common characteristics which tend to connect us more than they separate us. I believe that, for most of us, it's just a matter of degree as to whether we are somewhat more aware, or less aware, of bass localization.
In any event, I think that it may be past time to challenge the notion that most people can't localize bass at frequencies at or below about 80Hz, and consider the possibility that we potentially will be able localize 80Hz and lower frequencies, when we begin to investigate our subwoofer placements. As we test different subwoofer locations, we will certainly be able to discover whether bass localization is a factor for us.
FWIW, I think there is a reason why most people prefer to have at least one subwoofer on the front soundstage, and it may not be just aesthetics. In smaller rooms especially, where speakers and subs are closer together, a subwoofer on the front wall should help to prevent localization. But, I have seen specific instances where the best subwoofer positions, in terms of the measured frequency response, produced an unacceptable amount of bass localization. In those instances, as in other aspects of audio, individuals just have to pick the compromises which best suit their personal listening preferences.
Section III-E: LFE+Main:
There is a final setting, found in some AVR's (including Denon/Marantz) which some listeners may be tempted to try. It is called LFE+Main, or double-bass. It does literally double the bass, since it allows exactly the same bass content from the front two channels to be played by both the front speakers and by the subwoofers. With this setting, the subwoofers will continue to exclusively play all of the content in the LFE channel, but both they and the front speakers will duplicate the bass content in the regular channels. (Again, as long as a subwoofer is configured in an AVR system, it will always be the only transducer playing LFE content. That doesn't change with the LFE+Main setting.) It is called LFE+Main because the subwoofers still play LFE content, but now they also duplicate the bass content of the main speakers, instead of just handling the bass content that the main speakers don't play.
Some AVR makers added the feature for those HT owners who really didn't want to set their big front speakers to Small, with a crossover, but who still wanted to be able to utilize their subwoofers. But, as explained earlier in this section, the front speakers may not be able to play all of the bass frequencies in movies, or in some kinds of music, nearly as well as the subwoofers can. So, the front speakers may struggle with some low-bass content, causing audible distortion. And, when both speakers and subwoofers try to play the same frequencies, at the same time, the resulting sound quality can suffer in other ways, as well.
When we use LFE+Main, we are not redirecting the bass from our regular channels to our subwoofers as we are with typical bass-management. Instead, we are allowing the front channel content to be played by both the front speakers and the subwoofers. We can still set a crossover from the main speakers to the subwoofers if we wish, and the subwoofers will only duplicate content below that crossover. But, the front speakers and the subwoofers have very different output capabilities, very different frequency responses, and are EQed differently, so what typically results is considerable cancellation at some frequencies, and random peaking at other frequencies. But, the mid-bass frequencies might sound relatively louder, even if there were some resulting loss of clarity in the bass.
[To clarify what may sometimes be a little bit confusing, LFE+Main only operates when three conditions are met. First, the subwoofers must be shown (set) in the Speaker Configuration menu. Second, the front speakers have to be set to Large. Third, a setting of LFE+Main has to be used. That setting is found in the Bass menu, as an alternative to LFE. When LFE+Main is employed, any crossover still works for the front speakers. It is actually just a low-pass filter, at that point, which affects only the subwoofer. But, it is easier to continue to call it a crossover.
That crossover determines the frequency at which the subwoofers begin to duplicate the bass frequencies being played by the front speakers. So, for instance, if an 80Hz crossover is set, the subwoofers will softly start to begin operation an octave above 80Hz (as described in the discussion of crossovers) and will be at full-effect from 80Hz down. If a 60Hz crossover is set, the subwoofer will be in full operation from 60Hz down, and so on. Meanwhile, the front speakers, acting as full-range speakers, will play all of the bass content in the normal speaker channels down to the lowest limit of their individual capabilities. And, the subwoofers will continue to exclusively play all of the LFE content. The LFE part doesn't change as long as there are subwoofers configured in the AVR.]
It may be important to explain a little more about this idea of having transducers with different capabilities, and which are EQed differently, playing the same content. Audyssey and other systems of automated room EQ set filters for each channel independently. So for instance, the specific stereo content from the left front channel is subject to the EQ for that channel, and the specific content from the right front channel is treated the same way. Content below a crossover point for those speakers is EQed for all of the subs as a whole. And, HPF's for the speakers and LPF's for the subwoofers add gradual slopes which help to prevent excessive peaks in the response at bass frequencies. So, as one transducer (the LF, for instance) drops away, the sub picks up the volume.
But, when both speakers and subwoofers play exactly the same content, with front channels and the subwoofer(s), which have been EQed differently, the EQ that has been done can no longer be relied upon to prevent bass peaks and cancellation at random frequencies. In fact, it is fairly likely that the main speakers and subwoofer(s) will cancel each other at some frequencies, if they try to play the same frequencies.
The distortion that usually results from the LFE+Main setting may produce what is often referred to as one-note bass. That same muddy or boomy-sounding bass can often be heard when room EQ is not operating. Bass clarity occurs when every bass frequency can be heard more-or-less distinctly, without some frequencies peaking and other frequencies dipping or cancelling. LFE+Main rarely allows that clarity for individual bass frequencies to be heard.
To summarize the decision of whether or not to use LFE+Main, it may be helpful to compare it to the use of room correction, in general. People who are looking for bass clarity are more likely to find it when they use some system of room EQ. Turning-off room EQ will generally introduce some muddiness to the sound, as bass frequencies randomly cancel, or boominess as frequencies randomly peak. That may very well create the impression that there is more bass, as described in the first section, because we will typically hear some frequencies much more loudly than we will hear others.
That same thing may happen when we engage LFE+Main. It may increase the apparent quantity of the bass in exchange for some quality, in the form of bass clarity. There is nothing wrong with experimenting to determine which setting we prefer, and proceeding accordingly. But, I believe that it may be important to understand the potential trade-offs we are making, and that it may be important to listen objectively to our overall sound quality. Some AVR makers added the LFE+Main feature over the objections of the creators of Audyssey, and of others, who believed that it was contradictory to the fundamental concept of bass-management, and of room EQ as a means to enhance bass clarity.
One of the distinctions that I might personally make would be between the use of our full-range speakers for some types of music, and their use for other types of listening material. As noted earlier in this section, some people may prefer to use just their full-range speakers for music that doesn't have a lot of low-bass content. In that case, turning off the subwoofers in the Speaker configuration menu, and simply setting the front speakers to Large, would enable the listener to enjoy properly EQed bass, played entirely by his front speakers. And, even if someone prefers listening without the use of EQ, more clarity should be achievable with the front channels either playing full-range content by themselves or with the use of properly bass-managed subwoofers.
As a general rule, I think it's fair to say that, if we need to have our subwoofer(s) playing in order to have sufficient low-bass to begin with, then we are usually better off setting our speakers to Small, and bass-managing them with a crossover. Then, if we ever want even more bass, we can just boost our subwoofer volume to get it. That way, we can still benefit from having the right transducers playing the right content, for their specific capabilities, and we can benefit from having properly EQed transducers and from the improved bass clarity which should result. As with all of our settings, however, this is strictly a YMMV issue.
* The Guide continues in the next post, with Sections IV through VIII.
Regards,
Mike
Parts of the Guide may seem very thorough and detailed, but readers can choose the parts they wish to read at any one time. Over a period of time, however, most of the information contained in the Guide is likely to prove valuable for those who want to understand how things work, and for those who want to get the most from their HT/audio systems.
The Guide began on the Audyssey thread as a way to explain Audyssey's calibration process, and to provide guidance with respect to adding bass boosts after calibrations. But, over time, the Guide has expanded to encompass discussions of a wide-ranging series of topics involving our HT and audio systems, many of which have nothing to do with any system of room correction. The Table of Contents, and the Introduction to the Guide, provide an overview of the topics which are discussed.
* Audyssey remains the most commonly used system of room correction. Where Audyssey-related sections are involved, the basic principles of room/system interaction, of system calibration, and of room EQ, which are contained in the Guide, are believed to be generally applicable to other methods of HT calibration, and to other systems of automated room EQ.
With respect to the thread which follows the actual Guide, ideally this will be a thread where people will be comfortable discussing any HT calibration, subwoofer, or audio-related issues.
Cliff Notes:
(Tips For Getting Started With Your HT System)
Several people have suggested having some abbreviated HT calibration tips, so I have added a few simple tips prior to the actual Guide. Anyone wanting more detail on anything from room and speaker setup, to setting crossovers, to the differences between sealed and ported subs, to selecting and positioning subwoofers in a room, can find the pertinent information in the Guide itself. There are some similar abbreviated tips, to selecting subwoofers, at the beginning of Section VIII.
1. Try to position speakers and subwoofers strategically, doing a "subwoofer crawl" if necessary. Section I-B gives some good general advice on locating the speakers on the front soundstage. (The front three speakers will have the most impact on the overall sound quality for most of the frequency range.) You can Google how to do a sub crawl, or you can refer to Section VIII-E for instructions on how to perform the procedure.
2. Set the subwoofer phase at 0, and the low-pass filter (the LPF is sometimes labelled crossover) on the subwoofer at the maximum setting. Typically, the gain control on the sub (which is often labelled volume) should be set at or slightly below about the mid-way point. This varies! With some subs, the gain may need to be at 10:00 or 11:00 on an analogue dial. On others, such as some Monoprice subs, the gain may need to be at 3:00 or higher. During the level-matching process, which occurs at the first mic position, you will actually want the subwoofer(s) to be slightly above the volume that Audyssey is telling you to achieve, for reasons which are explained in #12. of the Cliff Notes. (On newer Denon/Marantz AVR's, that means that you will want your subwoofer volume to be slightly in the red zone.)
3. Allow your AVR to calibrate your audio system for you, using its automated routine. The first thing that will do is to calibrate your HT system to Dolby/THX Reference (which will correspond to 0.0 MV) , and it will provide a common basis for comparing listening levels and subwoofer boosts. That calibration process will insure that all of the channels are playing the same volume levels at the MLP (main listening position) and that all of the sounds will arrive at the same time. Those equal volume levels are also essential in order for the room correction process to occur. Room EQ only works to minimize peaks and dips in the sound if all channels are playing at the same volume level.
4. The way that Audyssey and other forms of auto-calibration work is that speaker levels and distances are set from the first microphone position, which is always defined by your AVR as the MLP. As noted above, that ensures that all channels play the same volume and that all sounds arrive at the same time. Subwoofers have internal processing, through their own internal amplifiers, which delays the arrival of the sound, and wireless subs have even more delay. So, their distance settings will not correspond exactly to their physical distance from the MLP. The distances will be greater than the physical separation from the MLP. This is normal, and the distances should be left as they are, unless there is some other specific reason to change them.
5. The second thing that the calibration will do is to set EQ filters, for all of the channels, to reduce some peaks and dips in the frequency response caused by the interaction of your transducers (speakers and subwoofers) with the room. Those random peaks and dips in volume, at different frequencies, interfere with the quality of the sound we hear. With subwoofers, boomy, one-note bass is often the result of a random peak. After room EQ, the bass may sound smoother, but correspondingly less impactful, until the subwoofer volume is increased. Increasing the subwoofer volume after calibrating is normal, as explained in Note 10.
6. It is important to understand that two different things are occurring during a calibration. The initial calibration process ensures that equal volume levels, from all of the channels, will arrive at the MLP at the same time. And it calibrates the audio system to a "Reference" standard. The room EQ process (that is also part of the calibration) sets filters, for all of the channels on an individual basis, in an effort to improve the overall sound quality in the room. To do that, it needs to start with equal volume levels. From those equal volume levels for each channel, room EQ will add or subtract volume at specific frequencies, to get as close as possible to the target volume of 75dB. Where room EQ is successful, a measured frequency response will show a somewhat flat line from the lower frequencies to the higher frequencies. (Once random peaks and dips are removed, listeners still have the option to tilt that somewhat flatter line toward their particular listening preferences.)
7. It is also important to emphasize that all channels are EQed individually, in relation to the room, and not in relation to each other. The speakers are not EQed with respect to each other or to the subwoofers. Each individual channel is EQed with respect to the room and the MLP. All subwoofers in a system are treated as a single channel, even if there are two Sub out's in the AVR. The two sub outs may allow the AVR to set separate volume levels and distances, based on the specific subwoofer positions in the room. Whether or not a particular AVR can set separate volume levels and distances will depend on the model. But, irrespective of whether there are separate sub outs, all subwoofers are EQed together, as if they were a single sub.
8. There is a model calibration procedure shown in Section I-C which may help you to achieve an optimum calibration. It has a diagram of potential microphone positions which seem to work well for many listeners. Starting with a good speaker setup, which is addressed in Section I-B, and a good calibration, can take some effort. But, doing those things can make an audible difference in the resulting sound.
9. As noted, when an AVR calibrates your audio system, all of your channels including your subwoofer(s) will typically be set to play the same volume at the main listening position (MLP), and you will be listening to Reference volumes when you are at a listening level of 0.0 MV (master volume). Most people probably listen at an average volume of about -15 MV to -20 MV. Individual listening volumes, however, can vary much more widely than that. (Your master volume is the only AVR setting that will be unchanged after an Audyssey calibration. It will still be wherever you had set it prior to running Audyssey.)
10. Since all of the channels are now playing equal volumes, and since we don't hear low-frequencies as well as other frequencies, after calibration most people will need to add more bass to their audio systems. That is particularly the case when we listen at below Reference volumes, where bass frequencies were designed to be in better equilibrium with frequencies in our normal hearing range of about 500Hz to 5,000Hz. As we drop below Reference (0.0 MV), bass frequencies drop-out of our hearing much more quickly than other frequencies do. It is fairly typical to add +3dB to +6dB of subwoofer boost on top of Audyssey's DEQ, and even more than that if DEQ is disabled.
11. We can compensate for the audible reduction in bass by turning-up the volume of the subwoofers, and how much volume to add is strictly a user preference issue. To state this in a different way, after an Audyssey calibration, very large subwoofers will be playing at exactly the same volume level as very small subwoofers would be, since all of the channels in a calibrated HT system are level-matched to play the same SPL at the main listening position. In order to use the greater available output of more powerful subwoofers, it is simply necessary to turn-up their volume after the calibration.
12. After running Audyssey, it is generally desirable to add most of your subwoofer volume increase with your subwoofer gain control, while not letting your AVR sub trim go above about -5, in order to avoid clipping the pre-out signal coming from your AVR. That is especially important at master volume levels above about -15. Typically, it is a good idea to raise the gain on the subwoofer high enough to achieve a trim level of about -9 to -11.5 during the initial level-matching process. (Gain and trim are inversely proportional during the level-matching process. Raise the sub's gain and the AVR trim goes down, and vice-versa.)
The lowest trim level that you should use with Denon/Marantz is -11.5, because the trim controls only go down to -12. If you are at a trim level of -12, you won't know whether your trim level actually should have been even lower than that. In some cases, it might have needed to be much lower than -12 if the trim levels could go that low.
An expedient way to set trim levels during a calibration is to just run three mic positions and tell Audyssey to calibrate. Then, you can look at the trim levels that Audyssey set. Once you get the AVR trim results you want, you can do a full 8-point calibration with the higher versions of Audyssey, or with 6 points for the lower versions. To make it even quicker, you can calibrate only for your front speakers and your sub(s). Just go into your Speaker Configuration menu to add or subtract channels when you use this approach. And, you can keep the microphone in the same spot for those three sets of sweeps. You are only setting trim levels to coincide at the MLP anyway, during the level-matching process.
It will probably take a volume level of about 78dB to 80dB, instead of Audyssey's default 75dB, to achieve low trim levels in about the -9 to -11.5 range. Where XT-32, with SubEQ is used, that will put the subwoofer volume in the 'red zone'. That's perfectly fine, just tell Audyssey to continue. After running Audyssey, we can conveniently raise the AVR trim to about -5 or -6, with our AVR remotes, and we can continue to increase the subwoofer gain if we want even more bass than that. Section II explains the best ways to use the subwoofer gain in some detail, and explains in greater detail why it is generally advisable to keep AVR subwoofer trim levels well in negative numbers.
* If a listener doesn't want to run Audyssey again, but has a subwoofer trim level that is higher than the Guide is recommending, the following procedure is perfectly acceptable. There is nothing wrong with simply lowering the AVR trim level, after a calibration, and then raising the subwoofer gain level to compensate for that. So for example, rather than doing a new calibration, a listener could lower the trim level to about -5 or -6 and then raise the subwoofer gain until the bass seemed loud enough. Ultimately, as with the master volume, everyone will just add or subtract bass in accordance with his own personal preferences.
13. It is important to understand that the amplifier connected to our subwoofer output in the AVR is not like the amplifier in our AVR which powers our speakers. The amplifier connected to our sub out is only intended to power on the subwoofer from Auto (Sleep) mode, and to make convenient incremental changes to our subwoofer volume. Our subwoofers have internal amplifiers which are needed for really serious increases in subwoofer volume. And, we can increase our volume with our gain controls as much as we want to, after a calibration, and use our AVR remote to make smaller up or down changes while keeping our AVR trim well in negative numbers.
There is a short article which explains this in more detail here:
Guide to Subwoofer Calibration and Bass Preferences
The Guide, which starts here in Post 1, and which continues in Post 2, is intended to be a general guide to Home Theater, HT calibration, and audio quality. Due to its roughly 250 page length, I have had to divide it into two posts. Sections I through III follow the Introduction in this post...
www.avsforum.com
14. Some Denon AVR's have a feature called Subwoofer Level Adjust. If a Denon AVR has an On/Off control for that feature, it should typically be turned off, and any volume adjustments should be made either with the subwoofer's gain control, with the subwoofer trim control in the Audio menu, or with the trim control in the Speaker: Manual: Test Tone area of the AVR. Again, in making any subwoofer trim adjustments, it is desirable to keep the trim levels at about -5 or lower. To add more subwoofer boost in excess of a -5 trim setting, it is always perfectly acceptable to use the gain control on the subwoofer. (You can keep track of your initial gain setting on an analogue dial by marking that hatch mark with a small piece of tape.)
15. It is important to understand that Audyssey will measure your speakers, at their specific positions in your room, and your AVR will set preliminary crossovers in accordance with its own programming. For instance, if a particular speaker pair, or center channel is capable of playing below 40Hz, at its specific position inside a room, and at the 75dB test tone, your AVR will set the speaker to "Large". Or a 40Hz or 60Hz crossover may be set, if speakers can't go quite low enough for an initial setting of Large. Those aren't actually recommendations! They are just observations, based on the measured response of your speakers. After an Audyssey calibration, it is advisable to reset crossovers as suggested below.
16. If there is a subwoofer in the system, speakers should typically be set to Small, and crossovers should typically be set at 80Hz or higher. (The 75dB test tone that our AVR's use isn't actually very loud. As we go up in volume, the speakers' low-frequency capabilities will degrade. Raising crossovers transfers more of the low-frequency demand to your more powerful subwoofer.) The LPF of LFE, in the AVR, should typically be set to 120Hz (which is usually the default setting in our AVR's). There are exceptions to these settings, which are explained in Section III, but this is typical best practice advice for starting-out with an HT system.
(As of October 2023, I have been informed that new Denon/Marantz AVR's no longer have a Large/Small setting. Listeners simply set crossovers for the various speakers.)
17. It is always acceptable (and often desirable) to raise crossovers from their initial calibration setting, but it is not generally desirable to lower them from wherever your AVR set them. Among other things, Audyssey will not be EQing speakers below the crossovers set during the calibration process. And, as noted above, 75dB is not very loud. As your speakers try to play peak volume levels greater than 75dB, they will roll-off faster and that will put more demand on them. It is usually better to let the subwoofers do the heavy lifting. Crossovers are explained in detail in Section III.
18. If you have Audyssey, you can add subwoofer boosts on top of Dynamic EQ, or turn off DEQ and add your own subwoofer boosts. Section V explains DEQ in detail. Experimenting with it on, and with it off, may be helpful. Turning it off will probably require you to add more subwoofer volume, but for some people, it may change the sound in a positive way. There are also RLO (Reference Level Offset) settings, which are associated with DEQ, and which may be helpful to moderate its effects.
19. You can turn Audyssey off, to hear how things sound without room EQ, and then turn it back on, without changing the room correction filters that it set for any of the channels. And, you can experiment with Audyssey Flat. Audyssey (Reference) and DEQ are always the default settings after a calibration. Audyssey Flat and/or DEQ off are user preference options, and both are explained in Section V.
20. After an Audyssey calibration, you can change any settings in your AVR without affecting the room correction filters that Audyssey sets. Changing AVR settings prior to running an Audyssey calibration will not be helpful. Audyssey is designed to ignore and override prior settings when it calibrates an audio system. The only setting that will remain unchanged after an Audyssey calibration is the master volume. But now, your AVR and your audio system will be calibrated to correspond to Dolby/THX Reference, when the MV is set to 0.0.
21. If you make a significant change to a room, such as moving a speaker or a subwoofer; changing to different speakers or subwoofers; adding new subwoofers or other channels; or adding room treatments or making significant furniture changes; you should recalibrate.
22. It may be important to recognize that virtually all settings, including your master volume level, are listener-preference settings. There is no universally correct way to listen to music, or to watch movies or TV shows. Even the use of room correction, in whatever form, is a user-preference feature. Some people prefer listening without room correction, or limiting its effect to just the lower frequencies. And, that same idea applies to all of the settings associated with room EQ, or with our AVR's in general. There are default settings that may help us to get started, and there are some best practice principles which we may want to follow. But, we will all define audio quality in slightly different ways, and we will all have slightly (or profoundly) different listening preferences. Informed experimentation can be the key to discover what we really like.
23. When you try different calibrations, or different settings, remember that you will ultimately have to trust your own judgment with respect to sound quality. The Guide, and other sources, can help to explain some audio theory, and how certain features work. Those same sources can make suggestions regarding general best practice principles, and can offer options for things that listeners can experiment with. But, in the end, everyone will have to decide for himself what he actually likes.
Each of us decides for himself how much, or how little, he wants to experiment with his audio system. Some of us may just be looking for a plug-and-play approach to our HT's. That is perfectly fine too! After all, the goal here is simply to please ourselves with respect to our entertainment hobby. And, for those of us who do want to experiment, each of us also decides when he is satisfied and wants to stop experimenting. It is not unusual to stop and just enjoy our audio systems for a while, and then to experiment again weeks, months, or even years later.
The way in which you experiment and listen can be important! Take your time! Try a particular setting for several days, unless you are absolutely sure that you don't like it, before trying a completely different one. Ideally, you want to let your hearing adjust to one sound quality before trying something different. What you don't want to do is to introduce several new variables all at once, because you won't be able to separate them, and you may not have really learned anything about your own listening preferences.
You also don't want to overload your own hearing, and your brain's response to what you are hearing, by trying to pack too many changes and too many concentrated listening sessions into too short a period of time. I referred to them as "concentrated" listening sessions, but that is really the wrong word. They really need to be as relaxed and natural listening sessions as you can make them.
Your brain will tell you what you do and don't like, if you relax and give it some time. You don't have to 'concentrate' to decide whether something tastes good or not, or how much seasoning you prefer, or whether you like a particular color. Concentrating on trying to hear specific things in your sound can actually be counterproductive. Just relax and enjoy the process of experimenting, and of making gradual, incremental improvements in your sound quality.
If you listen for two or three days, your hearing will adjust somewhat to that particular sound. Then, when you do try a different setting you will have an audio benchmark with which to compare any changes in the sound. You won't necessarily have to concentrate there either. Just let the listening sessions happen naturally. If there is no audible change in the sound, then you may not need to worry about that particular setting. If you feel that a new setting has a positive effect on the sound, make a note of that setting, and of your reaction to it. Keeping track of which settings work best for you, and why they seem to improve things, will help as you continue to experiment.
If a setting seems to have a negative effect, you might cross that one off your list right away, or perhaps return to it later. Above all, be patient and take your time. Impatience will only take you in circles! There are just too many different ways to achieve improved sound quality, and too many variables, for us to try to go too fast. But, if we are patient and systematic in our experimentation, almost everyone will get to a final result that is most appropriate for that particular listener. And, we can stop experimenting, and just enjoy our audio, anytime we choose.
____
Specific sections of the Guide deal with how to setup speakers and how to perform good automated calibrations. Other sections deal with setting crossovers, and selecting and positioning subwoofers. Most of the terms used in the Guide are defined in the first two sections. Readers can choose what sections to read for a particular purpose, although some of the information does build sequentially as you go along.
* The Guide is organized into the following major sections. Each section (and subsection) is hyperlinked so that readers can go directly to that part of the Guide by clicking on it in the Table of Contents. Individual sections or subsections can also be copied and pasted, by right clicking on them in the Table of Contents, for inclusion in a post. That will enable others to go directly to the subsection to which you are referring.
Note: With the transition to the new XenForo platform, the previous hyperlinks were lost. I have added new hyperlinks to allow internal navigation within the Guide, but they currently take us to a position about three lines down from where they should. The Forum Administrator has asked the Development Team to fix that glitch, but it appears that may never happen. In any event, it it still much quicker to use the blue hyperlinks, and then to scroll back up a couple of lines.
Table of Contents:
Introduction to the Guide:
Section I: Room/System Setup and Sound Quality:
Section I-A: The Frequency Range:
Section I-B: Distortion, Speaker Placement, and Room Treatments
Section I-C: Room EQ and Calibration Techniques
Section II: Audio System Calibration and Subwoofer Levels:
Section II-A: Audyssey Calibration And Dolby Reference
Section II-B: Why We Add Bass After Calibrations
Section II-C: Where And How To Add Bass
Section II-D: Master Volume Levels And Sub Boosts
Section II-E: Gain Settings And Maximum Sub Output
Section III: Setting Crossovers:
Section III-A: Crossovers From Speakers to Subwoofers:
Section III-B: Low Frequency Effects Channel:
Section III-C: Cascading Crossovers:
Section III-D: Bass Localization:
Section III-E: LFE+Main:
Section IV: Integrating Multiple Subwoofers:
Section IV-A: Setup and Calibration:
Section IV-B: Room EQ:
Section V: Audyssey Dynamic EQ and Dynamic Volume:
Section V-A: Dynamic EQ:
Section V-B: Tone controls and House Curves:
Section V-C: Dynamic Volume:
Section VI: Audyssey Thread History of Recommended Subwoofer Trim Settings:
Section VII: Bass Frequencies, Room Gain, and The Equal Loudness Contours:
Section VII-A: Bass Frequencies and Tactile Response:
Section VII-B: Room Gain:
Section VII-C: The Equal Loudness Contours:
Section VIII: Bass Preferences, and Subwoofer Selection and Placement:
Section VIII-A: Sealed Versus Ported Subwoofers:
Section VIII-B: Comparing Subwoofer Performance:
Section VIII-C: Selecting Single Versus Multiple Subwoofers:
Section VIII-D: Internet Direct Subwoofers:
Section VIII-E: Subwoofer Placement in a Room:
Introduction to the Guide:
The most commonly asked question on many AVR and room correction threads, and on a number of subwoofer owners' threads, involves subwoofer settings. People who have new audio or home theater (HT) systems, or who have upgraded and/or added subwoofers, are naturally anxious to be able to get the most from them. In addition, there is a fairly universal perception that bass volumes sound somewhat softer after running Audyssey, or YPAO, or other systems of automated calibration. And, people are frequently curious about whether that perception is normal, and if so, about the best way to increase their bass.
The Guide was originally written to explain why it may be perfectly normal to perceive bass levels as lower, after running Audyssey or other forms of automated calibration. And, it was written to explain the best ways to use a combination of subwoofer gain and AVR trim to make bass boosts. In attempting to address issues involving subwoofer boosts, however, I have found that it is also helpful to understand some of the basic principles of HT system calibration, and their relationship to Dolby Reference.
And that, in turn, has led to discussions of how we hear bass frequencies in relation to other frequencies, and of how our preferences influence our subwoofer selections and our subwoofer placements. As the Guide has continued to expand, I have also decided to try to address some fundamental issues of speaker placement and of how rooms influence the sound we hear. And, I have added some general suggestions on techniques to use during the Audyssey calibration process.
* Much of the information in the Guide may be helpful in understanding important audio and set-up issues, and will also be somewhat applicable to non-Audyssey systems of audio/HT calibration and automated room EQ.
It may be worth pointing out that we all like having some reassurance that we are operating our audio systems correctly, with the "correct" settings, and that we are getting the maximum benefit from them. I believe though, that the more that we understand some of the basic audio principles involved (which I certainly didn't when I first got into home theater) the more confidence we will be able to have in our own individual setting preferences, and in the resulting sound quality. As with almost everything in audio, sound quality can be very subjective, and it would be very difficult to identify a single set of "correct" settings which would please everyone.
Part of the key to developing a satisfactory audio system, in my opinion, is informed experimentation. AVS gives all of us an opportunity to share information with each other, so that we can enjoy our audio/HT systems more, and be more confident in the choices we make. And, that's really what the Guide is about--sharing information and, in some cases, speculation. I have learned a lot from writing it, and continue to do so as I try to add more detailed explanations. I hope that others will find it of benefit to them as well.
Sections I Through VIII:
There are eight major sections in the Guide, which begins in this post, and continues in Post 2. All but one of the eight sections are divided into multiple subsections, which cover a wide range of related material.
I. The first section starts with a description of the frequency range that we would be discussing in our home theater (HT) systems. Following that is an extensive subsection on system setup, and how rooms influence the sound quality we hear. It offers some advice on speaker placement and some fairly detailed discussion of room treatments. It also offers some calibration technique tips that may help people to achieve better results from an HT calibration.
II. The second section explains how Audyssey calibrates our audio systems. It is broken down into subsections which are labeled. The section explains the basic principles of how Audyssey works during the set-up process, and how it EQ's our audio systems. Many of the principles explained in Section II may also pertain to other systems of HT calibration and automated room EQ.
The second section also explains how audio systems are calibrated to a Dolby/THX Reference standard. It offers some best practice advice for getting the most from our subwoofers, and explains the relationships among subwoofer gain, AVR trim levels, and master volume levels. The section emphasizes the general desirability of keeping subwoofer trim levels in the negative range and using subwoofer gain to add sub boosts. And, it explains different ways to do that.
III. Since bass management is such an important component of all our audio systems, the third section explains some basic principles to consider in setting crossovers. The LFE channel, and something called bass localization, are discussed in some detail. And, a concept called Cascading Crossovers is introduced.
IV. The fourth section explains how Audyssey, and other systems of auto EQ, calibrate and EQ multiple subs. It also explains some of the difficulties that may occur when dissimilar subs are combined in an HT system. Phase cancellation which may occur between speakers and subwoofers, and which may also occur between subwoofers themselves, is discussed in this section.
V. The fifth section examines Audyssey's DynamicEQ (DEQ) and Dynamic Volume in some detail, and also compares and contrasts Audyssey Reference and Audyssey Flat. This section also discusses the use of bass and treble tone controls, and the development of Harman and more personalized house curves. That Section V-B has general applicability well beyond the use of Audyssey.
VI. The sixth section is a brief one that explains something of the Audyssey Thread history with respect to setting subwoofer trim levels, as the current advice is different from the advice in the much older Audyssey FAQ.
The last two sections have relatively little to do with Audyssey directly, or with room EQ in general, although there are some overlaps with room EQ. But, understanding some fundamental audio concepts, and especially some bass and subwoofer concepts, can enhance our ability to get the most from our audio systems.
VII. The seventh section is a longer one which explores the way that bass frequencies behave in a room, and which explores some of the general relationships among bass frequencies,including: tactile response, how room gain amplifies our bass, and how the way we hear and feel bass frequencies may influence the settings we use. Understanding those interrelationships is important! In that section, the Equal Loudness Contours, which illustrate how human hearing works, are also discussed in detail. Ideally, Sections VII and VIII will be read in conjunction.
VIII. The eighth and final section provides some fairly detailed guidance for people who are in the process of selecting subwoofers, and also provides some basic advice on positioning them within a room. Many people start threads on which subwoofer to buy, without having a good idea of what they are actually looking for, or how to distinguish among the options which people suggest to them. Section VIII will help with that. The section starts with some general rules to follow in selecting subwoofers--sort of like the Cliff Notes at the top of this page.
The five subsections in Section VIII go into considerable detail in describing differences between sealed and ported subwoofers; some different ways to compare subwoofer performance; the pros and cons of initially buying a single large sub, versus two smaller ones; and some descriptions and comparisons of some of the more popular ID (Internet Direct) subwoofer companies. Since subwoofer placement in the room is so important, a separate subsection is devoted to that.
[It is worth noting that the Audyssey FAQ, which is linked in my signature, and especially the Technical Addendum to the FAQ, have a wealth of additional information and explanation on some aspects of Audyssey which are not covered in this Guide. Interested readers are highly encouraged to read the FAQ, for both quick answers, and for some additional in-depth detail about Audyssey. However, wherever the Guide conflicts with the FAQ, the Guide presents more current and more accurate information, as explained in the Audyssey Thread History in Section VI.]
* REW: HT owners who are encountering specific problems with their frequency responses, or who wish to optimize their frequency responses (especially with multiple subwoofers), or who are simply curious about what is actually happening in their rooms, may wish to implement REW, which is a free download. Sometimes, people get very acceptable results from their automated room calibration systems, and sometimes the results are not satisfactory to them. In my opinion, this is an entirely personal decision which individual listeners will make for themselves.
In any event, measuring their frequency responses can tell those who are interested a lot about proper subwoofer positioning, set-up, and post-calibration adjustments. The use of REW will require a calibrated measurement microphone (a UMIK-1) and a computer (preferably a laptop) which can be connected to their AVR's or AVP's. Anyone interested in learning more about REW, and how to implement it, is encouraged to consult the following step-by-step guide by AVS member @AustinJerry.
https://www.dropbox.com/s/zdhq72a1puyyxpr/REW 101 HTS Current Version.pdf
There is also an AVS discussion thread which concentrates on the practical application of REW:
Simplified REW Setup and Use (USB Mic & HDMI...
The purpose of this thread is to explain how to both physically hook up the connections between your computer and AVR/Pre-Pro to get started with REW (Room EQ Wizard) and to share information on the proper use of REW including proper techniques for both measuring and interpreting graphs, what...
www.avsforum.com
Guide to Subwoofer Calibration and Bass Preferences
Section I: Room/System Setup and Sound Quality
There are a number of factors which can affect the sound quality in our rooms. Those factors include our speaker choices and their placement, distortion from the room itself, and the use of room treatments and automated room correction. There are a number of potential reference sources which can help us with our initial system setups, in terms of positioning our speakers, or with room treatments, or with specific room EQ calibration tips. But, I think that it would be worthwhile to try to address some basic concepts in this Guide, so that it can be a more general resource.
With that in mind, I would like to try to explain some basic concepts of system/room interaction and to offer some general advice on the relationships among the room, our system setups, and our sound quality. I would also like to offer some general tips on performing a successful calibration with room EQ.
It would probably be helpful to define some terms that are used in audio, and throughout the Guide. (Additional definitions and abbreviations are presented as they are used in individual sections.) I will start by using a good online definition of sound. Sound in air is made when air molecules vibrate, and move away from the vibrating source, in a pattern we refer to as sound waves. In our context, the vibrating source would be our transducers--our speakers and our subwoofers.
Sound pressure level (SPL) is a measurable quantity of sound volume. It is measured in decibels (dB). "Loudness" is not a measured quantity of volume; it is a perceived amount of sound. For instance, "That sounds really loud!" is a very different statement than "The SPL in the room is 100dB, as measured at the main listening position (MLP)." Loudness is a perception of how something sounds, while SPL is a measurable quantity of sound volume. The distinction between those two terms becomes very important when we are selecting our preferred listening volumes and our bass volumes.
The Subsections in Section I are as follows:
A: The Frequency Range
B: Distortion, Speaker Placement, and Room Treatments
C: Room EQ and Calibration Techniques
Section I-A: The Frequency Range:
Since we will be talking about various frequencies and how we hear them, throughout the Guide, I think that it would be helpful to begin with some explanation of how I would personally subdivide the frequencies that might be part of an audio/HT discussion. The lowest bass that can be meaningfully reproduced in an HT system is approximately 7Hz, although it is unlikely that most of us would be able to hear that low with complex content. Very high-frequencies can potentially be reproduced by modern tweeters, but the absolute upper limit of young healthy human hearing is 22,000Hz.
So, that 7Hz to 22KHz range is the one that we will be focusing on. But, how do we subdivide that range into divisions that facilitate a discussion of speakers, subwoofers, room treatments, and all of the other HT-related subjects that we may be interested in? That is the purpose of this subsection. I will begin this subsection by sharing two completely different graphs of the frequency range that we are discussing. We could find many others with a Google search.
The second graph is a little more complete than the first one, so in the discussion that follows, I will make some references to the Harman graph just above. It charts an actual in-room measurement of frequencies, with a rising-bass house curve added. (House curves are described in Section V-B. But, to briefly synopsize, we don't hear bass frequencies as well as other frequencies, so most people prefer to increase bass volumes, relative to those in our normal hearing range.)
I find discussions of bass frequencies very interesting, and also very confusing. I have researched this topic on numerous occasions, and have always found completely different ways to subdivide the frequency range of human hearing. There is particular disagreement as to the upper limit of what is a "bass" frequency, as opposed to a mid-range frequency. (There is also disagreement about what exactly is a mid-range frequency.)
I think that if we look for graphs online, we will find some agreement that ULF (ultra low-frequency) bass is <20Hz, although many frequency graphs stop at 20Hz. And, we will frequently see the range between ULF and 50Hz defined as the low-bass range. (FWIW, I think that 30Hz also has some special significance, as that is where we start to have trouble distinguishing between sound and physical vibrations.) Where mid-bass is defined at all, it will typically start at 50Hz (although it's sometimes 60Hz) and it often extends up to about 100Hz. (I personally prefer to use 120Hz, for reasons that are explained below.) But, that still leaves the upper-bass range, and that dividing line is all over the place.
Here is part of the problem as I see it. Everyone who is attempting to define bass frequencies is approaching the definition from a slightly (or dramatically) different perspective. Some of the people who are defining bass, mid-range and high-frequencies are musicians, and they tend to approach the issue from the standpoint of the musical instruments themselves. For instance, a 4-string upright bass has fundamental frequencies that range from a low of 40Hz to a high of 400Hz. So, since the upright bass is specifically designed to be a bass (and a deep-bass) instrument, some musicians might define bass frequencies as extending to about 400Hz.
Alternatively, some of the people who offer subdivisions of bass frequencies are recording mixers. And, by and large, I think that most of them define bass frequencies as extending to somewhere between 200Hz and 300Hz. It is shown as 250Hz in the Harman graph. In some respects, that 200Hz to 300Hz definition of the upper limit of bass frequencies seems even more arbitrary than the musical one, with each graph using a different dividing line.
Still another definition of bass frequencies comes from some audio engineers and HT hobbyists, who think in terms of the Schroeder (transition) frequency in a room. That is the frequency where low-frequencies become standing waves inside a room. Most of those definitions put an upper limit on bass frequencies of about 200Hz, because that seems to be about the upper limit of that transition frequency even in very small rooms.
I say that it "seems" to be the upper limit, because if you attempt to use any of the online calculators to determine the transition frequency in a room, you will get widely divergent results from the different calculators. If our definition of what is an upper limit of bass frequencies is dependent on the room size and construction, the specific room geometry, and the reverberation time within the room, then this may actually be the least useful definition of all. The definition of what is the upper limit of bass frequencies will vary with every room under discussion.
Another definition of the upper limit of what is a bass frequency comes to us from three-way speaker designers, who design their woofers to have upper limits, and their mid-range drivers to have lower limits, and who then create internal crossovers between the two. Those crossovers between woofers (bass drivers) and mid-range drivers is typically somewhere between about 300Hz and 400Hz, although some three-way speakers are probably outside of that range on either end. (It is also important to note that a woofer in a three-way speaker has to be able to play a little above the crossover. So, if the crossover is at 300-400Hz, the woofer has to be able to play frequencies up to at least 500Hz and higher.)
Here is an example of a definition that attempts to bridge several approaches. This one comes from a speaker designer and the bold emphasis is his. I will let individual readers make their own sense of this one:
"The bass frequencies cover 20 to 1,000 hertz, while treble covers 1,000 to 20,000 hertz and mid-range overlaps from 300 to 3,000 hertz. Mid-bass range is approximately 140 to 400 hertz. A mid-bass woofer is a speaker specifically designed to handle this sound frequency."
I will give that one points for originality, if for nothing else. Confused yet? I certainly am! Given the wide disparity in definitions, and the lack of an apparent logical basis for most of them, I decided several years ago to come-up with my own divisions for use in the Guide. At least that way I can explain my reasons for selecting the division of frequencies that I use. The divisions I use have evolved a bit from where I originally started them, as I have learned more, and thought through things a little differently. I have also wanted to achieve better consistency in my methodology, and I believe that my current approach does that.
I don't claim that my division of frequencies is "correct" in any universal sense, or even in any specific use of that term. It is simply a reasonably logical way of dividing the frequency range, that I think may be useful in discussing audio, and its application to our home theaters.
Dividing the Frequency Range:
Starting with bass, I would define the upper limit of the bass range in the following way. To me, it is approximately 500Hz (~480Hz according to the Equal Loudness Contours), where our perception of loudness starts to change. Below that frequency, we require more volume to hear sounds in equilibrium with those in our normal hearing range. That seems to me like a logical place to say that bass frequencies are starting. And, it's only a little higher than a couple of the other definitions that we saw.
Supporting that upper limit is the fact that frequencies below about that frequency begin to radiate more omnidirectionally, rather than in a more directional fashion. That means that the bass frequencies are leaving a speaker cabinet in all directions, rather than just coming from the general direction of the speaker cone. As noted, most speaker makers also cross from mid-range drivers to woofers in about the 300-400Hz range, although some cross a little lower or higher than that. So, if we established 500Hz as the upper limit of bass frequencies, I think that we would be in the right general ballpark for most HT discussions.
(I could also support a division for bass that was a little lower than 500Hz; perhaps in the 300-400Hz range. But, for the purposes of the Guide, I prefer the logical consistency of using the Equal Loudness Contours, and the way our hearing changes at about 500Hz, for this division.)
Mid-bass frequencies may also be a little difficult to define, and we don't always see that range specified in frequency graphs, such as in the two examples above. But, in HT, it seems to be a pretty important range, which comes-up all the time when people are selecting, configuring, and EQing subwoofers. I like defining mid-bass as the range from about 50Hz to 120Hz. That's a fairly small range, but it has several things to recommend it. First, for most people, that seems to be the average range where chest punch sensations are felt, and most people already associate those chest punch sensations with mid-bass frequencies.
I emphasize the word "most" here, because our perception of chest punch probably follows a bell curve, with some people outside the norm at both ends of the curve. Several studies have reached similar conclusions on that approximate 50Hz to 120Hz range, with one blind study determining that 63Hz was the frequency where most of the participants felt chest punch most strongly. At least two ID sub makers provide a pre-programmed PEQ boost centered on that specific 63Hz frequency.
To me, another reason that the 50Hz lower limit makes some sense for mid-bass, is because that is about the frequency where we can often observe a difference in the performance of ported subs and sealed subs, where sealed subs are starting to roll-off compared to ported subs, which maintain linearity at that frequency. There can be exceptions to that generalization, especially with the very largest and most powerful sealed subs. But FWIW, I like defining frequency ranges that correspond to some pragmatic HT considerations.
Using 120Hz as the upper end of the mid-bass range also makes some sense to me, since that is the upper limit of the .1 low-frequency effects channel (LFE) which was specifically intended to be played by subwoofers in the original Dolby/THX standards. The low-pass filter for the LFE channel is not a brick wall, and some sounds creep-in above 120Hz. But, sound mixers are primarily trying to amplify specific bass content in that channel only up to 120Hz. (Or perhaps, only up to 80Hz with respect to the most meaningful bass content.) In any event, that 120Hz low-frequency effects cutoff point seems like a logical separation between mid-bass and upper-bass.
There can also be multiple ways to define what constitutes low-bass. For instance, as noted earlier, many graphs which describe the frequency range don't even consider frequencies under 20Hz. So, low-bass would just be anything under about 50-60Hz. If we define low-bass as about an octave-and-a-half below the 50Hz limit that we set for mid-bass, and ULF as <20Hz, we have the following relatively proportional divisions, which go from the lowest frequencies to the highest frequencies:
* 7Hz to 20Hz: ULF, which covers the frequencies below 20Hz. That would be about 1 1/2 potentially meaningful octaves, using the 8-note per octave scale, where each doubling of frequency is one octave. (For example, the frequencies between 10,000Hz and 20,000Hz would still just be one octave, consisting of 8 distinct notes.)
* 20Hz to 50Hz: Low-bass would be about the 1 1/2 octave range from 20Hz to 50Hz.
* 50Hz to 120Hz: Mid-bass would be the roughly 1 1/2 octave range from 50Hz to 120Hz (125Hz would make it an exact octave-and-a-half, but that's splitting hairs.)
* 120Hz to 500Hz: Upper-bass would be the 2-octave range from 120Hz to approximately 500Hz. Below 500Hz is where our perception of equal loudness starts to change.
* 500Hz to 5,000Hz: I would define the mid-range frequencies as covering the frequency range from about 500Hz to 5,000Hz. That is just a little more than 3 octaves, which seems about right for the frequency range where our hearing is the strongest. Most speaker designers seem to cross their mid-range drivers to their tweeters at just about 2,500Hz to 3,000Hz. In fact, mid-range compensation, found in some audio curves such as Audyssey's default Reference curve, is centered on 2,500Hz. That -3dB reduction in SPL at 2,500Hz is based on the original "BBC dip", which was designed to improve crossover blending from mid-range drivers to tweeters.
Since mid-range drivers need to play a little above a crossover of around 2,500-3,000Hz, it makes some sense to define the upper limit of the mid-range as 5,000Hz, in order to provide some cushion for that. Part of the reason for using 5,000Hz, as a dividing line for mid and high-frequencies, is also for the sake of consistency with what we used for the upper limit of bass. Our normal hearing range, where all frequencies sound equal in loudness, is the range from 500Hz to 5,000Hz. According to the Equal Loudness Contours, our perception of loudness changes at 5,000Hz, with frequencies above that sounding a little softer, just as they start to do below 500Hz. So, it makes a certain amount of sense that the "middle range" would correspond to our normal hearing range of 500Hz to 5,000Hz.
* 5,000Hz to 22,000Hz: Treble or high-frequencies. Based on my current thinking, high-frequencies would start at about 5,000Hz, and continue all the way up to 22KHz, which is the extreme upper limit of young and healthy human hearing. (Most of the people reading this, including the person writing it, probably can't hear much above about 12KHz anymore, if we can hear even that high. But, that's another story.) That frequency range would be approximately 2 octaves.
Most music-related definitions of high-frequencies draw a distinction between fundamental frequencies, which only extend up to about 6,000Hz for almost all musical instruments, and harmonics (one and two octave overtones) of those frequencies, which add 'brilliance' to the sound. There is a graphic illustration of the range of musical instruments in Section I-B. They then generally subdivide the mid-range category, extending up to about 4,000Hz, into three separate divisions: low-mid, mid-mid, and high-mid, as illustrated in the Harman graph.
Between 4,000Hz and 6,000Hz, in that same Harman graph, is something called "Presence". Everything above 6,000Hz is then considered "Brilliance". I don't personally find any of the low-mid, mid-mid, and upper-mid divisions to be particularly useful for audio/speaker/HT purposes. Nor, do I find the terms "Presence" and "Brilliance" especially helpful for HT use.
And, since some musical instruments do play fundamental frequencies above 6,000Hz, that first division at 4,000Hz, and the second one at 6,000Hz, seem somewhat arbitrary to me. I have some idea of what is meant by "presence" in musical terms, but the term doesn't carry enough intrinsic meaning to be very helpful in our HT discussions. I do think that the use of the term "brilliance" can be helpful, in a general descriptive sense, for musical instruments. And, I have also sometimes used the term "bright" or "brilliant" to describe the high-frequency sound of some tweeters, or of high-frequencies inside a relatively untreated room. But, for HT discussions, at least, it may be a little too ambiguous a term to constitute a meaningful frequency division.
(FWIW, I think that the difference between 5,000Hz and 6,000Hz is pretty inconsequential when we realize that there are only 8 distinct notes in the octave between 5,000Hz and 10,000Hz. So, if someone else wanted to define high-frequencies as starting at about 4,000Hz, or at 6,000Hz, I certainly wouldn't have a problem with that. I would hope though, that there would be some specified basis for the definition, as there is here.)
In any event, I like the fact that this overall subdivision of frequencies seems to have some inherent logic and proportionality. And, for audio/HT purposes, I find that definitions of frequency ranges are more helpful, when they somewhat align with the way that our speakers and subwoofers work, and with the way that we actually hear different frequencies. To me, if we just think of the way that a three-way speaker works, we gain some insight into how to subdivide frequencies for HT discussion purposes. Woofers play bass frequencies (with some overlap above their internal crossovers), the mid-range drivers play mid-range frequencies (with some overlap both above and below a crossover), and the tweeters start playing softly below the crossover to the mid-range driver, and then play all of the treble frequencies.
Once again though, I think this points to the real nature of the problem in these discussions. Are we coming at our divisions of the frequency range from the standpoint of the operating range of musical instruments, or from the perspective of recording mixers, or are we thinking of subwoofer/speaker performance in an HT? For HT purposes, I find the subwoofer/speaker perspective (combined with the way that our hearing actually works) to be the most useful way to talk about divisions within the frequency range of about 7Hz to 22KHz. But as with almost any audio/HT-related issue, there can be other legitimate viewpoints.
Section I-B: Distortion, Speaker Placement, and Room Treatments
What we hear, when we listen to a recording, is heavily influenced by the room itself. To begin the discussion in this section, it may be helpful to talk a little about some factors that influence what we hear, and to talk about room-related distortion in a general way. There are much more sophisticated explanations of distortion that readers can investigate, but I want to try to cover some very basic concepts. First, distortion can be defined as "any alteration in the waveform of an audio signal." (All sound is composed of frequencies--sound waves--so any alteration in the waveform can affect what we hear.)
In practice, "distortion" is sort of a catchall term for nearly anything which adversely affects perceived sound quality. Some audio sources we listen to may have compression in the source itself, or distortion in the specific recording. Speaker, subwoofer, and AVR makers go to great lengths to minimize distortion, and particularly Total Harmonic Distortion (THD) in their products. Anyone interested in understanding more about audio distortion, from audio sources and transducers, is encouraged to consult more authoritative articles such as this one:
Blog - How Much Distortion Can We Hear With Music? | Axiom Audio
Since frequency response (FR) is mentioned so often in audio discussions, and in the Guide, it is also worthwhile to define what is meant by a "smooth", or "even", or "flat" frequency response. A smooth, even, or flat frequency response would be one in which no frequencies are playing significantly louder or softer than other frequencies. (We will often refer to peaks and dips in the FR.) In music, and in movies, some sounds will be deliberately emphasized over other sounds--such as a trumpet solo in a music recording, for instance, or low-bass sequences in movies. That would be inherent to the recording and not dictated by the room. However, the room will inevitably influence the evenness of the FR in unpredictable ways, causing some frequencies to sound louder or softer than others. We can remediate that potentially uneven FR with good placement of our transducers, with room treatments, and with room EQ.
Some of us may be deliberately creating a house curve to change the interaction of our recordings and our room, in some specific direction (such as having less treble, or having more bass). But, most of us would not typically want the room itself to arbitrarily dictate which frequencies are louder or softer than others, in a way that has nothing to do with the original intent of a recording. Where a frequency response is reasonably "flat" or "smooth", all frequencies are playing at approximately the same volume, at a particular point in space, unless we, or the recording, dictate otherwise. The selected point in space where everything is more-or-less in equilibrium is the main listening position (MLP). As we move away from the MLP, the relative smoothness of the FR may change, as some frequencies become slightly louder or softer, compared to others.
1. Room Distortion:
If we drive a particular speaker, or subwoofer, or AVR too hard, we may introduce distortion into our signal chain. But, even if we are playing content at moderate volumes, once we play an audio recording, from any kind of transducer (a speaker or subwoofer which converts electrical energy into sound) located inside a listening room, the interaction with the room itself influences the sound that we hear. Not all influences are negative, as the room may influence the sound we hear in a number of positive ways. But, not all room influences are positive, either. It is the negative influences caused by interaction with a room, which we may characterize as distortion, that I am going to try to address in this brief discussion.
Section VII explains in more detail the difference between frequencies below the Schroeder, or transition frequency in a room, and frequencies above that point. Briefly though, below the transition point in a room, which is typically about 200Hz or lower (depending on the size of the room), some bass sound waves slow down when they encounter a room boundary, and they collect as "standing waves". Some bass frequencies are amplified, and some are cancelled, as a result of the interaction of those standing waves with room boundaries. Room boundaries would typically consist of four walls, the ceiling, and the floor, although there could be more or fewer than four walls in a given room.
Overall, room boundaries reinforce the bass SPL, but the amplification/cancellation of individual frequencies can be problematical. And, they are inevitable whenever a bass transducer is placed inside a room. (We can mitigate that somewhat with good speaker/subwoofer placement, and with the proper placement and integration of multiple subwoofers.) Higher frequencies, especially above about 300Hz, behave differently than the lower frequencies do. They aren't amplified or cancelled in the way that low-frequencies are. But, when direct sounds and reflected sounds from those higher frequencies arrive too closely together at a listening position, they can cause distortion in what we hear.
Mid-range and treble frequencies may not really reinforce the measurable SPL in a room, but they certainly may sound as if they do, especially in very lively-sounding rooms. In a room with a lot of hard surfaces, some of what we are perceiving as loudness is probably distortion. (I am using distortion in this case as a catch-all term. In a small room, sound waves bounce back-and-forth for a longer period of time than they would in a much larger room. The more bare surfaces there are in the room, the more prolonged reverberation we will have.) It is easy to understand this if we think of the difference between singing in a shower, with bare walls and a tiny space for the sound to reverberate in, and singing at the same volume level in a normal size bedroom.
We typically perceive high-frequency distortion as being louder than undistorted sound. In fact, we may perceive it as being much louder. And, most of us don't seem to tolerate higher frequency distortion as well as we do lower frequency distortion. In fact, for some people, high-frequency distortion can be a little painful. The term "ear fatigue" is sometimes used to describe hearing high-frequency distortion during a listening session.
The words we use to describe different types of distortion tell us something about our discomfort. For instance, we may describe high-frequency distortion as screechy or piercing (which suggests a somewhat painful sound), where we might describe low-frequency distortion as muddy or boomy (simply suggesting a lack of clarity). The words we use to describe the different types of distortion are significant, and the difference in the way we hear distortion at different frequencies can be important. Understanding differences in the way we hear some frequencies, and learning something about the distortion that may accompany those frequencies, can be helpful as we try to improve the sound quality in our rooms.
[It may be important to understand that home theater (HT) rooms of about 20,000^3 (^3 is an abbreviation for cubic feet) or smaller, sound perceptually louder than commercial cinemas, which are always much larger than even very big HT's. A number of audio experts estimate that our HT's may sound anywhere from about +5 to +9dB louder than the same SPL would seem when played in a commercial cinema. The +5dB to +9dB range of difference varies with the size of the room, with very small rooms (<1500^3) being as much as +9dB louder sounding than the same volume would be if played in a commercial movie theater. The following table illustrates the relationship between room size and perceived loudness, based on an assumed volume level of 85dB in a commercial theater. As you can see, in a room smaller than about 1,500^3, only about 76dB would be required to equal the apparent loudness of a commercial theater playing at 85dB:
As noted earlier, some of that louder-seeming sound might potentially be due to distortion. But, even in a heavily treated room, the smaller size of the room would appear to amplify the loudness level we would perceive. I have heard room treatment consultants estimate that a well-treated room might sound up to about -3dB less loud, than it did prior to room treatment. I suspect that the degree of loudness attenuation might depend somewhat on the size of the room. For instance, a very small room might benefit even more than -3dB from extensive room treatments. But, it would still sound significantly louder than the same SPL played in a commercial theater.]
In order to understand how rooms can create distortion, it is important to understand something about how sound waves behave in a room. For the purpose of the discussions in the Guide, I am defining bass frequencies as those frequencies below 500Hz. As noted previously, I would probably define mid-range frequencies as the range from 500Hz to about 5,000Hz. As a practical matter, frequencies from about 2,500Hz or 3,000Hz and up would typically be played by tweeters, as they have to play slightly below the internal crossover in a speaker.
This may be a good place to illustrate a description of what musical instruments play what frequencies. Understanding that will help to correlate what is written in this Guide, and elsewhere, with what we are hearing with acoustic (non-electronically enhanced) instruments. The following graphic is one popular interpretation of the frequency range of musical instruments. The red horizontal lines represent fundamental frequencies, and the yellow lines represent harmonics (overtones) of those frequencies, which add brilliance or sparkle to the fundamental sounds that we hear.
Sound waves are vibrations, with different lengths and different vibration speeds. (They vibrate--moving back-and-forth, sort of like a slinky toy, as they travel through the air.) They all travel at the speed of sound (which is approximately 1 millisecond per foot at sea level), but how they behave in a room, and how we perceive them when they reach us, makes all the difference in what we hear.
As frequencies get higher, the vibrations become much shorter in length, and they oscillate (move or travel back and forth between two points) much faster. They also tend to travel in straight lines unless absorbed, or redirected, by contact with some surface. Longer frequencies oscillate more slowly, and they tend to bend rather than always simply ricocheting in a straight line. They also go right through solid objects in a way that higher frequencies cannot. Those differences can have a direct bearing on what we hear in a room. And, the differences in sound wave lengths are very significant.
For example, a 10Hz frequency is only 10 cycles per second (meaning that it vibrates 10 times per second), and the wavelength is about 112' long. By contrast, a 100Hz frequency is 100 cycles per second, and the wavelength is now only about 11' long. As we go up in frequency, the wavelengths get shorter and shorter. A 1,000Hz frequency is 1,000 cycles per second, and the wavelength is just a fraction over a foot long. By the time we are up to 10KHz, and 10,000 cycles per second, the wavelength is only a little more than an inch long. The table linked below illustrates frequency length:
Frequency - Wavelength - Period Chart
Where low-frequencies oscillate relatively slowly, bending and pooling in corners (where any two or more room boundaries meet), the frequencies above the transition point (which is usually above about 200Hz or so in most rooms) oscillate increasingly faster, and bounce around in a room, reflecting off of hard surfaces like a billiard ball bouncing off the cushions on a billiard table, until they run out of energy from friction. The billiard ball analogy actually falls a little short, as all frequencies, but especially mid and high-frequency sound waves, ricochet off of all six surfaces in a room (the four walls, the floor, and the ceiling) until they run out of energy, or are absorbed by something. (An even better analogy would be a handball court.) Higher-frequencies are absorbed more easily than low-frequencies, and they run out of energy faster due to heat exchange with the air and with the surfaces they touch, than is the case with low-frequencies.
Sound waves may also bounce off of other hard surfaces, such as table tops or other furniture. Those reflected sound waves which don't arrive very close in time to the first sounds to reach our ears, may be somewhat ignored by our brains. We typically hear the first-arriving sounds as being louder, and we concentrate more on those sounds. Our brains can typically filter-out, or separate, the later arriving sounds, if the sounds are somewhat delayed, in order to concentrate on those louder, first-arriving sounds. It's sort of like unconsciously tuning-out the conversation at an adjoining table, in a restaurant, if we are interested in what our dinner companion is saying.
But, higher frequency sounds arriving within about 6 milliseconds (ms) or so of each other are often associated with distortion, because our brains can't distinguish them as clearly from the first arriving sounds. (As noted earlier, sound travels at approximately 1ms per foot at sea level, and very slightly slower than that at high altitudes due to cooler temperatures.) I suspect that how much we might notice distortion, from sounds arriving too close together, can vary somewhat depending on the individual, and on both the frequency and the SPL of the sounds. For instance, distortion usually becomes more noticeable at higher volume levels.
In any event, sounds arriving close behind the first-arriving sounds from our speakers, within approximately that 6ms window or so, may distort the sound we hear. In a best case scenario, early reflections may simply make some sounds seem louder or more three-dimensional. But, they may also be perceived as contributing to a harsher sound with some mid-range frequencies or to a more strident or piercing sound with high-frequencies.
To summarize, a reflected mid-range or high-frequency sound from a wall or other hard surface, which arrives at our ears at almost the same time as the direct sound from a speaker, may create a type of distortion, as our brains will have difficulty combining the separate close-arriving sounds into a single coherent sound. When that happens, we may experience ear fatigue, or we may simply suffer a loss of clarity in the sound we hear.
Distortion may, or may not, be particularly noticeable, depending on whether we are actually accustomed to hearing undistorted sound from our audio systems. This is an important point! The expression that "we don't know what we don't know" could be expanded to "we may not hear what we don't know to listen for". A personal example of that is given later, involving employees at a high-end audio store who were always accustomed to hearing distortion in their store's listening room, and who consequently didn't know that the music recordings they played for customers in order to audition expensive speakers and amps weren't really supposed to sound that way.
And, there can also be a component of individuality involved. Not everyone is going to hear exactly the same frequencies, in exactly the same way, and our listening preferences may vary as well. As volumes increase however, excessive early reflections may cause distortion which is painfully shrill or harsh to some people. As noted, "ear fatigue" or "listening fatigue" are terms which may often relate to some form of mid-range or high-frequency distortion.
Our initial system setup can contribute to the relative clarity and fidelity of the sound that we hear in a room. Some suggestions regarding subwoofer placement in a room are addressed in the last section of the Guide. In this first section, though, I would like to concentrate on the speakers, and on their relationship to the room and to the listening positions. I would particularly like to concentrate on the speakers located along the front soundstage, as they carry the majority of the content in a movie or music recording, and will have a much greater influence on our overall sound quality.
(There are numerous on-line guides illustrating placement objectives for surround and overhead speakers, but I believe that insufficient attention may often be paid to the front soundstage which is so critical for both music listening and movie/TV viewing.)
In the context of the discussion of how sound behaves in a room, anyone who is interested in how specific frequencies sound, and how the room may influence our perception of sound, may find the following video helpful. In the video, individual frequencies are played, starting at 20Hz and continuing to 20,000Hz. The test should be conducted only in stereo, and the sound should always be coming from the center. Where the sound seems to come more strongly from one ear than from the other, the room your speakers are located in is influencing the sound, as room reflections from one speaker or the other momentarily seem to dominate the sound. It's a great illustration of how the room influences what we hear.
Something else that the video demonstrates is how low some low-bass frequencies may actually sound to us. Most of us probably believe that some frequencies we hear are lower than they really are. Getting a feel for how low in frequency an 80Hz, or 60Hz, or 40Hz tone actually is can be quite revealing. That is why a consistent theme in the later sections on bass is that we feel as much as hear frequencies below about 30Hz. And, it can take a fair bit of SPL to hear and feel those very low-frequencies, in conjunction with the sounds in our more normal hearing range, which may be accompanying them.
Pure tones, like those in the video, and the complex sounds which we hear in music and in movies, are very different. Complex sounds may consist of multiple frequencies, each containing both fundamental frequencies and harmonics (overtones) of those frequencies. Most of the time, the very low-frequencies are simply adding some bass weight to sounds, rather than standing-out as distinct sounds themselves, although there can be some exceptions to this with some low-bass content.
2. Front Speakers:
To continue the discussion of room distortion, let's start with the front speakers. Many speakers require some distance from the wall behind them in order to achieve good sound quality. Electrostatic speakers may require even more separation from a wall as they don't radiate sound in the same fashion as direct-firing speakers. Bipole speakers are designed to radiate sound both forwards and backwards toward walls. And dipole speakers are designed to radiate most of the sound backwards toward walls, in order to create a more diffuse sound field.
Most speakers are direct firing--meaning that the sound primarily travels in the direction they point toward. With direct-firing speakers, the sound leaves the drivers, in the front of the speaker, in a cone-shaped cluster of sound waves. But, even with direct-firing speakers, other sound waves travel backwards, through the speaker cabinet, to reflect off the wall behind the speakers. (Rear-ported speakers may also require some wall clearance to operate properly. About 4" is a good rule-of-thumb for port clearance.) Bass frequencies will radiate omnidirectionally, from all sides of the cabinet.
With front speakers something to be aware of is called SBIR. It stands for Speaker Boundary Interference Response. It may or may not ever be an issue to specific listeners, but it is worth mentioning. If a speaker is too close to a wall, there may be some boundary interference from that wall. Low-frequencies radiating backwards, or toward a sidewall, would strike the wall and reach the listening position less directly then the waves traveling in a straight line from the speaker. That, in turn, could cause them to be out-of-phase with the direct wavelengths.
Where that happens, some bass frequencies would be amplified, while others were cancelled, causing peaks and dips in the overall frequency response. In some cases, the dips will be significant. But, in some respects, the peaks will be even worse. Our brains will concentrate on first arriving sounds and will to some extent disregard later arriving sounds. But, if direct and indirect sounds arrive too close together, they will create distortion in the sound. Improving the effects of SBIR can not only allow us to hear frequencies that might previously have been cancelled, but it will improve the overall clarity of the sound.
Moving front speakers away from walls can help to prevent SBIR. Placing approximately 2" to 4" thick broadband acoustic panels behind speakers which are close to walls, can also help to prevent boundary interference, where it is believed to be a problem. A good way to determine whether or not SBIR is a problem is simply to experiment with moving speakers closer to, and further from walls, and listening to the results. You are simply listening for the placement where the sound is the clearest, prior to running room correction.
The following video is excellent in explaining what SBIR is and how to deal with it:
To return to the general discussion of speakers, several things are happening when a speaker plays inside an enclosed space, such as an audio/HT room. First, the direct sound is arriving at the listening position. Second, sound waves which bounced off the wall behind the speaker, are arriving just a little later at the listening position. Third, sound waves which bounce off the side-walls are arriving at the listening position just a little behind the direct sound. And fourth, sound waves which bounce off the floor and ceiling (and to some extent off the wall behind the listener) are arriving at the listening position just a little behind the direct sound. And, all of those sounds will arrive very close together. Our brains are remarkably adept at sorting-out all of those very early and later-arriving sound waves into a single coherent sound. But, when the sounds arrive too close together, they can't, and we hear distortion.
Opinions vary as to how to address distortion caused by the speakers and the room. For instance, speaker makers typically try to walk a fine line with respect to having drivers with wide horizontal dispersion, in order to provide a wider-sounding soundstage and in order to maintain the same SPL, off-axis, as they do when they are pointed straight ahead. But, if they get too wide, early reflections from side-walls may become more of a problem. Again, it is helpful to remember that sound waves above about 500Hz leave a typical direct-firing speaker in a cone shape, with the narrow portion of the cone closest to the speaker, and the mouth of the cone getting wider as the sound waves move further away from the speaker.
I think that most speaker makers try to achieve about 30 degrees of horizontal off-axis dispersion, at a distance of about 3 to 4 meters. That would represent the typical listening distance for most tower speakers, or for larger bookshelf speakers. Vertical dispersion is usually more restricted. I believe that about 15 degrees is typical in order to avoid too much reflection from floors. It is important to note that some very large speakers are intended for relatively large rooms, and for longer listening distances--up to 5m or more. Knowing this can be important, because large tower speakers, which are too close to a listening position, may not be able to achieve as good inherent sound quality, without the intervention of room EQ, as speakers which are designed for that typical listening distance referred to above.
[This issue of having large speakers, designed for greater listening distances, may be worth pursuing a little further. If we think of a typical three-way tower speaker, the speakers (woofer, mid-range, and tweeter) are arranged vertically, with each driver spaced a prescribed distance from the other drivers, so that they will voice appropriately at a particular distance. (There is more to it than just the driver spacing, but this conveys the general idea.) That means that the drivers are designed to all play at the same volume at a designated distance. The distance is usually a range. As noted above, I believe that 3 to 4 meters is typical for a tower speaker, or for a large bookshelf speaker.
Smaller bookshelf speakers, however, may be designed for a listening distance of only about 2m. For instance, My two-way Audioengine A5+ speakers only have a tweeter and a 5" mid-range/woofer driver. They are designed to be used primarily as desktop computer speakers, without an external amplifier or AVR. And, they are self-powered for that purpose. They are intended to be separated by anywhere from about 3' to 6' from each other, and they are designed to voice most effectively at about those same distances. To facilitate that, the drivers in each speaker are in a relatively short vertical cabinet, and the drivers are spaced closely together. Spaced appropriately, they image extremely well in my desktop system, with good sound separation and a good phantom center in stereo.
A very large tower speaker which is designed to be employed in a large room, on the other hand, may have speakers which will play at equal volumes at a distance of 5-6 meters or more. I actually have a pair of very large (18^3) speakers like that. They were specifically designed to be separated from each other by about 20' and to each be about that same distance from the main listening position. They were always intended for larger listening venues, and they don't really sound as good in a smaller room.
Automated room correction can help the various drivers, within a vertical cabinet, to play the same volumes at the calibrated listening distance. (It will amplify some frequencies to play at the same sound level as other frequencies.) But, starting with speakers which are actually sized reasonably appropriately, for the listening distance, is still a good idea. And, this is a concept which is sometimes overlooked when people select speakers. They may select large tower speakers, which voiced well in an audition room at an audio store, but which don't sound quite as good, at their own listening distance, in a smaller room.
As noted above, automated room EQ can help somewhat to alleviate issues with our listening distance from our speakers, because EQ filters (control points) can be set at various frequencies to insure that all of the frequencies reach the listener at the same volume level. That is slightly different from the distance setting which governs the arrival time of the overall sound from a speaker.
But, when speakers are situated at approximately the appropriate distance from a listener, and at the appropriate height with respect to the listener, all of the drivers in a well-designed speaker should start-off playing at about the same inherent volume level, at least prior to room influences. That makes it much easier to achieve a smooth frequency response, and good sound quality, even before automated room EQ sets filters to improve the speakers' specific interaction with the room. It should also be noted that not everyone will have room correction to start with, or will choose to use room correction for higher frequencies. More on that later.]
Finding the sweet spot for a particular pair of speakers, in a particular room, usually requires some experimentation. For instance, if a speaker is pointed too far away from the listening position, there may be an adverse impact on the high-frequencies (which are typically very directional), and it may result in an increase in early reflections from the side-walls. So, determining how far to keep speakers away from the wall behind them, and from the wall beside them, and how much to toe-in the speakers toward the listening position, is something that may require some trial-and-error. Ideally, listeners will try to perfect the sound quality of their speakers, in their rooms, before running room EQ. That is because the better the native response is to start with, the less that room correction is likely to interfere with the quality of the sound. (There is more on this subject in the following subsection on calibration tips.)
The old stereo rule-of-thumb about trying to keep a roughly equilateral triangle with respect to the two front speakers and the MLP (main listening position) is still a pretty good starting point, in order to determine the best separation for the front speakers. That doesn't relate directly to distortion, but it does relate to the ability to easily generate good stereo and HT imaging, and it may also relate to the ability to create a realistic soundstage width, which extends beyond the screen for HT, or beyond the speakers for stereo. So, soundstage width, and the overall benign or malign influence of wall reflections, may also be factors when trying to decide how close front speakers should be to side-walls.
Another factor involving the placement of the front speakers concerns the early reflections from side-walls referred to above. It was suggested earlier that most sound waves leave a direct-firing driver in a cone shape, and that most speaker makers try to get about 30 degrees of off-axis horizontal dispersion. Depending on how close a speaker is to a side-wall, and how much (or how little) the speaker is toed-in toward the MLP, the more that early reflections from side-walls may be an issue. That might be especially true for speakers where speaker manufacturers specifically suggest pointing speakers straight ahead, or even toeing them away slightly, as that could result in more side-wall reflection.
Speaker positioning, with respect to early reflections from adjacent walls, is entirely a matter of personal preference. Some manufacturers of acoustic panels, and some audio experts and AVS members, believe quite strongly in always treating side-walls for early reflections. On the other hand, listening tests conducted by Dr. Todd Welti, Dr. Floyd Toole, and others, indicate that most people prefer the additional ambient sound provided by untreated side-walls.
According to those listening tests, most people hear a wider and more realistic sounding soundstage when those side-wall first reflection points are either untreated, or at least, less heavily absorbed. Something which disperses (scatters) sound waves rather than absorbing them might be helpful, in some cases. That could include something like a bookcase, or an acoustic panel designed to act as a diffuser. The idea is simply to create an irregularity at early reflection points, so that sound waves won't ricochet from the walls in predictable, straight-line, patterns. As noted, for some listeners, a wider soundstage may make the sound seem to come from beyond the sides of the speakers.
Listeners need to make their own determinations of what they prefer in their specific rooms. People wanting to experiment with their preference for more ambient sound from side-walls, versus treating those early reflections, can try putting some sort of fabric on a wall temporarily, to test whether or not they like the result. Determining where to put a piece of test fabric, such as a blanket or a folded towel, or something with a surface irregularity, is fairly easy.
One person sits at the MLP, while another person holds a mirror on the side-wall. Where the seated person can see the speaker in the mirror is the first reflection point for that wall. And, that demonstrates where some type of temporary absorbing or diffusing material can go. If a mirror test results in subjectively better overall sound quality, for a particular listener, he now knows where to install nicer and more permanent acoustical treatments.
3. Center Channel:
The center channel (CC) is an extremely important component in an HT system, as it carries so much of the content in movies and TV programs. And, the placement of the center channel can be very important for dialogue clarity in both movies and TV. The front speakers are typically designed so that the tweeters are more-or-less at the ear level of a seated person. That is typically the case with both tower speakers, and with bookshelf speakers on stands. But, most center channels (except for those placed behind an acoustically transparent screen) are necessarily either below or above the display (or screen). And, they are typically horizontal speakers in order to facilitate those placement options.
[That horizontal placement necessarily means that most center channels, in mixed use rooms, will sit on some sort of cabinet. If possible, the speaker should not be enclosed within a cabinet, in order to avoid the same boundary interference that was discussed with the front speakers. The shelf a center channel sits on, however, is rarely mentioned as a potential issue with respect to SBIR.]
When CC's are placed above or below a screen, higher frequencies (which are highly directional) are not beaming straight at most listeners. As noted earlier, vertical dispersion, from both mid-range speakers and from tweeters, is usually restricted in order to avoid too many early reflections from floors. Consequently, when the speakers don't point more directly at listeners, that can interfere with overall sound clarity, and particularly with dialogue intelligibility.
(This is just speculation on my part, but I believe that where a center channel is not pointing directly toward us, we are still likely to be able to hear most of the lower frequencies in the human vocal range, and that would include most of the vowels. But, if the higher frequencies are attenuated, due to being so directional and pointing away from us, or if they are distorted by interference with a cabinet, or by early reflections, some sounds may be harder to hear. That would include some of the consonants, such as "B", "C", "D", "G", "P", "T", "V", and "Z", especially at the beginning or end of words.
Sounds from those consonants, and from other fricative sounds such as "F" and "S", which are produced at the front of the mouth, involve higher frequencies. Without them, the intelligibility of individual words can be lost. And, that may make it harder to understand what is really being said. That might especially be the case with softer or more rapid speech, where foreign accents are involved, or where background sounds are partially masking dialogue.)
It is advisable to make sure that CC's are pointed as much toward ear level as possible. Where a CC is above the screen, a shim placed under the back edge of the speaker will point it down more toward ear level. (Based on AVS Forum experience, however, it is harder to get clear sound from a center channel which is placed above a screen and pointing downward. And, it also a little harder to maintain the illusion that voices are actually coming from the characters on the screen.)
Where a CC is located below a screen, a shim placed under the front edge of the speaker will point it up toward ear level. Getting the center channel pointed more directly toward ear level can sometimes make a substantial difference with respect to audio clarity. And, when a center channel points directly toward our ears, our brains will usually adapt even more readily to the illusion that the speech we hear is coming from the mouth of the speaker on the screen.
Another placement issue that can affect both overall audio clarity and dialogue intelligibility, is the extent to which a CC is recessed on a shelf or within a cabinet. The potential for SBIR was mentioned just above, but there are also other issues with a CC placed inside a cabinet. As sound waves emerge in a cone shape, high-frequencies reflect off the top and bottom (or sides within a cabinet) of a shelf, causing what's called comb-filtering effects. That's a very jagged frequency response, with lots of little peaks and valleys, at higher frequencies, that interfere with sound clarity.
The result of those very early reflections from the shelf, arriving so close in time to the direct sound waves, can create a distorted sound and can especially interfere with our ability to hear dialogue clearly. Pulling the center channel forward so that the front of the speaker protrudes about an inch clear of the shelf or cabinet will help to prevent comb-filtering.
[Where it is possible to do so, it is always better acoustically to not enclose speakers (or subwoofers) inside a cabinet. It may be counterintuitive to think that enclosing a wood cabinet speaker, inside another wood cabinet, would negatively impact the sound. We may think that the cabinet would act to reinforce the sound and add more bass. It should reinforce the sound, from the standpoint of having sound waves bouncing around inside the display cabinet, or whatever the enclosure is. But, those sound waves which bounced off the sides, top, bottom, and back of that cabinet will reach our ears just enough later (or out-of-phase with the direct sound) to interfere with the clarity of the direct sound from the speakers. If possible, speakers should always be at the forward edge of a shelf, and if they can be taken completely out of cabinets, that is advisable.]
Irrespective of good center channel positioning, it can sometimes be a little difficult to hear dialogue in movies clearly. There are a number of factors that could contribute to a loss of clarity in general, or to dialogue intelligibility in particular. Sometimes, center channels are simply a little too weak to keep up with the other speakers in a system, or with the subwoofers. The CC may be the most important speaker in an HT system, as it is involved in playing at least about 80% of the content in a movie, and nearly all of the dialogue.
Heavy bass boosts can sometimes contribute to difficulty in hearing dialogue, especially where crossovers of 100Hz or higher are used for the CC. When a higher crossover is used, strong bass boosts may make voices sound deeper, thicker, and harder to understand. In that case, a better CC, which can utilize an 80Hz or 90Hz crossover, may be helpful. (A procedure known as cascading crossovers can also be helpful with dialogue clarity. It is explained in Section III-C.)
The use of DEQ (an Audyssey program) may also affect dialogue intelligibility, in some cases, as it slightly boosts the bass in all of the channels, including the center channel. And, it boosts the surround channels by about 1db for every -5db of master volume. The louder sounding surround channels could make it harder to hear the CC. (DEQ is discussed in detail in a later section.) Even without the action of DEQ, ambient noises, music, and special effects in a movie or TV soundtrack, may make sounds from the center channel harder to hear, and voices harder to understand. Many people prefer to boost the volume of their center channel a little, depending on the specific program, or on their specific circumstances.
4. Early and Late Reflections, and Room Treatments:
I have mentioned early reflections from side-walls that may involve the front speakers, and from cabinets and shelves that involve the center channel. There are other areas of the room that involve both the front speakers and the CC. If the floor between the front soundstage and the listening position is a hard surface, such as concrete, tile, or wood, there are almost certainly going to be some early reflections which can interfere somewhat with sound quality. In that case, an area rug in front of the listening position can be very helpful. Typically, people will use a foam rubber pad under the rug to prevent slipping, and to partially attenuate reflections from frequencies above about 1000Hz.
Coffee tables directly in front of a sofa can also be a source of early reflections, as high-frequencies bounce off the hard surface and reach our ears just behind the direct sound from the speakers. Any softening influences that can be added to the table can help. Even scattering some magazines on a table can help to disperse high-frequencies, so that they don't reflect directly toward a listening position. Remember that higher frequencies leave the speaker in a cluster, and they tend to travel in a straight line, so scattering (diffusing) them can also help to reduce the distortion we may hear.
Two other areas of a room can also be particularly problematical. First, as noted earlier, speakers radiate some sound waves backwards. If they are fairly close to a wall (perhaps within a foot or two), it can be advisable to put something such as an acoustic panel, or an oil painting, or decorative fabric behind the speaker. That will prevent early reflections from the wall from interfering with the clarity of the sound reaching our ears. If a dense acoustic (rockwool or fiberglass) panel is used, in a 1" or 2" thickness, it will typically only affect frequencies above about 240-300Hz. Foam rubber treatments, such as the egg crate versions shown in a video below, may not absorb frequencies below about 1000Hz, although they can also act as diffusers. But, enough of them in a room can have a significant effect, especially with high and mid-range frequencies.
If a listening position, such as a sofa, is within about 2' or 3' of a wall, it may be very helpful to put something on the wall behind the sofa. Ideally, something like an oil painting, or a tapestry, or an acoustic panel, would be used in order to disperse or absorb sound waves. Otherwise, those mostly high-frequency sound waves would bounce off the wall behind us, and into our ears, just a few milliseconds behind the direct-arriving sound from our speakers. Reflections from that back wall could also interfere with our audio clarity. Aside from hard surfaces, such as a floor or table, directly in front of a listening position, this might be the most important early reflection point when a listening position is near a wall.
This idea of reflections from front and back walls, and from floors and table tops, is probably worth expanding on a little bit. If there are reflections from a cabinet top holding a center channel, or from the floor, or from a table located within a few feet of a listening position, the result may be the same. That is because, in each case, the sound would arrive at a listener's ears only a few milliseconds behind the first arriving sound. (Again, sound travels at ~1ms per foot at sea level.)
The same thing would apply to reflections from a wall behind a speaker, or from a wall just behind the listener. In both of those cases, the reflected sounds would arrive just a few milliseconds behind the first arriving sounds. For instance, let's assume that the main speakers are pulled out 2' from the front wall. Some mid and high-frequency sound waves would travel backwards through the cabinet to reflect off the wall behind them. Their round trip of 4' to strike the wall and ricochet toward the listening position, would put their sound arriving ~4ms behind the sound waves that came from the drivers firing directly toward the listening position. That 4ms second difference could be just enough to create some audible distortion.
Distortion could also occur from sound waves reflecting from a wall behind a chair or a couch. If the couch were within a couple of feet of a wall, then sound waves would go past the couch, hitting the wall and ricocheting back toward the listening position. That could also create distortion, as the sound waves would still arrive at our ears well within that approximately 6ms window referred to earlier. (Again, when our brains can easily separate sounds, the first arriving sounds will seem louder, and we will tend to ignore the later arriving ones.)
I have observed that when someone is able to move the couch out about 3' or so from a wall, the problem is much less noticeable. In that case, the reflected sounds would be arriving a full 6ms or more behind the first arriving sounds. And, in some cases, that is just enough for the first arriving sounds to mask the later arriving ones. When in doubt, putting some acoustical material on the wall behind a couch is a good idea.
Examples of decorative products which can absorb higher frequencies, and which can be strategically placed in appropriate locations in a room, are all over the Internet. Here is an example of one such product:
AcousticART Panels with Custom Graphics and Images – Printed with YOUR OWN Photos or Art
* Reducing Reverberation (Ringing):
Most of what I have addressed so far would be classified as early reflections. But, as noted in the first part of this section, frequencies continue to bounce around a room until they simply run out of gas, and that can cause audible problems too. As noted earlier, in the discussion about distortion, we may particularly notice when higher frequencies do too much of that. And, the presence of many bare or hard surfaces in a room allows higher frequencies to continue to ricochet around the room for a much longer period of time. When a relatively bare room has a lot of excess sound energy in it, the result can sometimes be perceived as shrill or harsh sounding. That shrillness can often make it difficult to tolerate very high volume levels.
The same thing happens with low-frequencies in an untreated room, but the sonic effect of the longer time that it takes those frequencies to decay is very different. As noted earlier, low-frequency distortion is often described as boomy or muddy, where higher frequency distortion may sometimes be described as sounding too harsh or shrill. Both low and high-frequency distortion indicate a lack of clarity in the sound, although our reaction to them may be slightly different. I will start with a more complete discussion of higher frequencies as they are often the most noticeable and objectionable.
Listeners sometimes note becoming fatigued by their high-frequencies. Some speakers may be more likely to create that sensation than others, and some listeners may be more susceptible to that sensation than others. But, in either case, a very bare room will usually exacerbate the sensation of brightness or shrillness in an audio system. I have been in a listening room in a "high-end" audio store that sounded like fingernails on a chalkboard to me, at anything above a low volume level, simply due to all of the bare surfaces in the room. Some listeners, such as the employees at that audio store, can become so acclimated to a particular sound that they may not even notice the overall distortion they are hearing.
Some audio magazines may contribute to an impression that bare room surfaces are appropriate for "audiophiles" by publishing photographs of audio systems in very bare rooms. I have seen photographs of very high-end speakers, with equally expensive amplifiers, in rooms with all bare surfaces, including polished hardwood floors. The rooms were attractive and the speakers looked very nice in those photographs, but I would personally not have enjoyed listening to music in those rooms.
Whether or not to add any rugs, or drapes, or other softening influences to a room, is certainly a personal decision, and sometimes there may be important WAF (wife approval factor) considerations, as well. But, in general, bare rooms will promote distortion, which may be especially noticeable at mid and high-frequencies. It is important to note that ringing (meaning a longer decay rate of sounds within a room) is not something which will be visible from a graph of the frequency response of an audio system.
A waterfall graph (obtained through the use of REW) can illustrate where longer decay rates are occurring, but they can be both difficult to interpret and hard to correlate to what we actually hear. The physical reverberation time of sounds can be measured, as an R60 value, with the right equipment and technical understanding. But just as with a waterfall, we still have to correlate the reverberation time to our own listening preferences.
Where a listener lacks measuring equipment, but suspects that the bare surfaces in his room may be contributing to excessive high-frequency energy, a simple handclap test will help to determine how much of a factor the room itself is playing in the sound he hears. Standing at the listening position and simply clapping your hands sharply will tell you a lot about the room. If the sound of the handclap is prolonged, the reverberation we are hearing is called slap echo, or flutter echo, or simply ringing. Ringing interferes with our ability to hear individual sounds distinctly, because louder or first-arriving sounds linger for too long, drowning out harmonics of the initial sound, or the more subtle sounds which follow.
A very good example of this is an instrument such as a chime or a cymbal. There are subtle harmonics from the fundamental sounds produced by those instruments (and many others) which go way up in frequency. (A fundamental sound or frequency is a single note, but as stated earlier, all musical notes have harmonics which go up in frequency, at a somewhat reduced volume.) When the room itself exhibits ringing, we won't hear those harmonics as clearly, or at all. We will just hear the first loud fundamental sound, and the ringing in the room will drown-out the more subtle harmonics, which would otherwise cause the sound of the chime to linger in the air for just a moment, with the correct timbre, as the recording intended for it to.
This idea may seem a little counterintuitive at first, but what I am saying is that if the room itself has ringing (prolonged reverberation which would be almost like an echo in a larger space), then the natural timbre of an instrument, which is captured in a good audio recording, may be eclipsed by the inherent ringing of the room itself. And, the musical, or other sound, will be cut-off somewhat abruptly, and harmonics of that sound may be supplanted by room reverberations of the fundamental sound, which continue to reverberate in the room.
Other good examples of instruments to which this might somewhat subtly apply are the piano or other string instruments, or wind instruments such as the flute or clarinet. I suspect that most of us won't even be aware of what we might be missing unless we experiment--comparing very lively-sounding rooms, to less lively-sounding rooms, and listening for the differences.
The YouTube video I am linking below has a terrific illustration of the ringing phenomenon I have been describing. In this instance, the instrument is a snare drum, playing in a room with all hard surfaces, oblique angles, and no furniture. This is a deliberately extreme example, as most rooms will have a little more favorable geometry, some furniture, and some other softening influences.
So, in most cases, the ringing effect would be much more subtle than what we hear in the video. And, in some cases we might not really notice it without a before-and-after comparison. It would also typically take much less effort to reduce the ringing in a normal room, and we might not want to deaden the room quite as much as is illustrated in the video. (This is an issue that will be explored a little more below.) But, the video dramatically makes the point of what excessive reverberation in a room can do to distort mid and high-frequencies.
GIK Acoustics has a short article which builds on the example of the untreated room by illustrating differences at lower frequencies as well as higher frequencies. And, they show before-and-after waterfall graphs, along with a short audio track demonstrating the audible difference that adding full-range bass traps and other acoustical treatments can make:
Understanding Decay Time and Waterfall Graphs - GIK Acoustics
Decay time waterfall graphs and frequency response charts are tools to measure the acoustics of a room. This article explains what these graphs and charts mean.
www.gikacoustics.com
Individuals who want to enhance the clarity of frequencies above the transition point may wish to investigate whether early and late reflections are a problem in their rooms. If so, simply adding some softening influences to the room may greatly enhance the quality and clarity of the sound in the room. Room treatments can be subtle and don't necessarily have to involve the use of acoustic panels. Area rugs, thick drapes, bookshelves filled with books, and other softening, absorbing, or diffusing influences can also be very effective with higher frequencies.
The YouTube video illustrated above above provides a great example of what excessive room reverberation sounds like with upper-bass and mid-range frequencies, using a snare drum. In my experience, any sound which is intended to linger a little is compromised by excessive reverberation. The more dominant sound lingers instead, and the subtle harmonics of the original sound are lost. High-frequency sounds, which have a certain degree of inherent brilliance, due to their harmonics, may become especially distorted when reverberation times are high. And, once heard and recognized for what it is, that type of distortion is very hard to unhear. I definitely notice it in rooms with relatively poor acoustics.
When ringing in a room has been addressed, a handclap test reveals a single fairly sharp sound, with relatively little lingering quality. And that, in turn, allows the full frequency range of individual instruments, and of individual sounds to be heard, without the room itself getting in the way. To what extent this is an issue in a particular room, for a particular listener, is another YMMV decision. As a general rule, it is probably better to add acoustical treatments and/or softening influences to a room gradually, listening as we go. Concentrating on the specific areas of the room, mentioned earlier in this section, would also be a good idea.
How much overall reverberation, or ringing, we prefer will undoubtedly vary from individual to individual. Most of us would probably want to shoot for a reverberation time (RT60) of somewhere between about 0.2 and 0.7 seconds, for our HT's. But, that is a very wide range in terms of perceived sound quality. (RT60 is an industry standard for the time it takes for a sound inside a room to decay by 60db.)
I have seen both 0.4 seconds (400ms) and 0.7 seconds (700ms) described as optimum HT targets from two different home theater designers. I believe that part of what makes the different recommendations confusing is both the goals and the preferences of the individual designers. A longer reverberation time (in the .5 to .7 range) would perhaps provide better music fidelity, but might also make movie dialogue slightly harder to distinguish. A shorter reverberation time (in the .2 to .4 range) might provide more dialogue clarity, but with a slight sacrifice in liveliness for music. Everything is a compromise!
But, that's where personal experimentation and personal preference come in. Finding the "optimal" compromise for a particular HT room is an extremely individualistic exercise, and that's why it can be helpful to go slowly and to listen carefully when adding room treatments. Another aspect of this is probably important to point-out. Room size and room geometry matter with respect to reverberation time. Larger rooms can support longer reverberation times than smaller rooms can, and that's one reason that the ideal reverberation time is shown as a fairly broad range.
Remember the illustration that showed the correlation between room size and perceived loudness. In that graphic, a 1500^3 room sounded +4dB louder at the same volume level (SPL) than a 5,000^3 room did. A small room in the 1,000-2,000^3 range would probably lend itself much better to a .2 second reverberation time than a 5,000-6,000^3 room would. The room geometry and the location of the listening area probably also make a difference.
The worst type of room, for reverberation, would probably be a small square room, such as a 12' by 12' by 8' room (1152^3). And, the reverberation issue would be intensified near the center of the room. (It could probably also be intensified, especially for bass frequencies, by lighter construction materials and a suspended wood floor.) It is likely that getting a reverberation time much closer to .2 seconds would be desirable in that room. A longer rectangular room (such as 12' by 20') or a room with an irregular geometry, could probably support a little longer reverberation time and still deliver comparable sound quality. And, a much larger room, such as mine at 6,000^3 with an irregular geometry, can support a reverberation time around .5 seconds or higher.
One final factor that could be worth mentioning is listening volume. Someone who never listens with a master volume above -20 or -15, could probably get by with less room treatment, and a higher R60 value, than someone who listens at master volume levels above -10. So that could also be worth considering as room treatments are added. In any event, it is clear that the decision as to how much room treatment to do, in order to achieve our personal audio goals, is an extremely personal decision.
The article which follows explains reverberation time in both simple, and more technical terms, as we follow the associated links.
Room & Building Acoustics
To get a sense of how we might determine the approximate reverberation time in our rooms, with a handclap test, the following series of short videos may be very helpful. In the first brief example, the RT is probably about 0.2 seconds (200ms). That room might feel very dead or dry acoustically--perhaps similar to a recording studio. The next one demonstrates an RT of 1.0 second. I would personally consider that room to be a little too lively for music listening and for HT. Somewhere between the first two examples (perhaps around 0.3 to 0.6 seconds) is where most people will probably want to be. Short of measuring our rooms with appropriately implemented acoustical analyzers, performing a handclap test should give us a good general sense of how we are doing when we add acoustical treatment to our rooms.
How to get a feeling for RT60 value
If we go slowly with our acoustical treatments, adding softening influences gradually, and paying attention to our own perceptions of sound quality, there should be no risk at all of over-treating a room in a way that compromises desirable ambient sounds or helpful reverberations. It is probable that different individuals will prefer slightly more, or less, room treatment, so two different people could probably address the same room in somewhat different ways. The main thing is to become aware of the ways that our rooms may be negatively affecting our sound quality, so that we may better please ourselves with the aesthetic/acoustic choices that we make.
A rule-of thumb which always made sense to me, for audio in general, was that a room which sounds comfortable for normal conversation is likely to sound comfortable for listening to recorded music, and perhaps also for watching movies. That is because, just as with normal conversation, the room itself wouldn't be adding or subtracting too much to the sound for music listening, or for movie watching.
A room which is so lacking in reverberation, that it isn't very comfortable for hearing normal conversation, at normal speaking levels, may also be adversely affecting the normal sounding timbre of voices, and especially of higher frequencies. (Remember the earlier musical terms that described "presence" and "brilliance" in frequencies above about 4,000Hz.) The same dullness that we might hear with normal conversation may be expected to carry over to what we would hear with recorded music.
Alternatively, a room which is so reverberant that voices sound abnormally loud or somewhat distorted for conversation, probably won't sound very good for music, or for movies, either. Musical instruments, playing higher frequencies, and vocals, may be particularly affected, as we will miss out on subtle harmonics. And, some sounds may become relatively shrill. Movie dialogue can also be especially affected by excessive room reverberation, as the center channel producing the dialogue is already competing with several other sound sources, even without excessive room reverberation to exacerbate the problem. So, in HT systems, excessive room reverberation may mask dialogue clarity in movies. In all cases, though, this is a YMMV issue, which individuals will need to resolve to their own satisfaction. Sometimes aesthetics, or even simple inertia, will trump acoustics, and that's fine too.
** Bass Traps:
For frequencies below about 250Hz or 300Hz, bass traps can be very helpful in reducing the overall decay rate in a room. And, as in the audio example that GIK used in an earlier link, reducing the bass decay rate in a room could make the bass sound clearer and less boomy. But, it takes thick and dense acoustic panels to affect frequencies below about 120Hz. Bass traps are acoustic panels which are specifically designed to absorb some longer bass wavelengths. As explained in more detail in Section VII, low-frequencies behave more sluggishly than higher frequencies do, pooling in areas where any two (or more) room surfaces meet.
For that reason, they are referred to as "standing waves". When standing waves collide at room boundaries, the low-frequency wavelengths either reinforce each other, causing peaks at certain frequencies, or they cancel each other, creating dips or deep nulls. Peaks at random frequencies can create a boomy, one-note-bass effect. Dips and deep nulls make it difficult, or impossible, to hear some bass frequencies.
As some of those excess standing waves are collected by the acoustic traps, random peaks and valleys in the frequency response are reduced, allowing other low-frequency sound waves to be heard more clearly. Boomy sounding bass is often the result of a loudness peak, at a particular frequency, that obscures other low-frequency sounds. On the other hand, bass traps can often allow bass to sound both clearer and louder than it did before, when significant cancellation was occurring. As demonstrated in that GIK audio track, bass boominess and an overall lack of clarity can also be attributable to the bass decay rate in a room. And, bass traps can help to remediate that.
Unlike other room treatments, used to reduce ringing at higher frequencies, it is probably more difficult to go too far in reducing the reverberation time of frequencies below about 200-300Hz. I believe that would especially be the case in a smaller room, involving relatively lighter construction and/or a suspended wood floor. Longer bass reverberation times in a small room could really obscure other frequencies, and sound extremely boomy to some people. So, I think that most people would probably be safe in adding as much bass trapping as they wanted to, although it is likely that not all of the bass traps would need to be full-range. Even there though, there could be some degree of personal preference involved in selecting the preferred amount of bass reverberation in a room. Going slowly, and adding treatments gradually, is still probably advisable.
Bass traps are typically at least 4" thick and are typically made of compressed fiberglass or rockwool. The ones with which I am most familiar have a plywood backing with a large hole cutout in the plywood. The hole works something like a Helmholtz resonator to attenuate low-frequencies. Then, the entire panel is covered with fabric. The best results are obtained when the thicker panels can stand-out away from the wall, with a gap between the panel and the wall behind them. I believe that using about a 4" gap behind a bass trap is usually recommended for very low-frequency absorption. The combination of the thicker acoustic panel, and the air space behind them can make bass traps somewhat effective down to about 50Hz or 60Hz, where the panel sitting flush against a wall, might only be effective down to about 120Hz.
Thick bass traps (4" to 6" thick) can be placed anywhere, but are frequently recommended for placement in corners, as that is nearly always a location where low-frequencies collect. The large corner traps are typically wedge-shaped to fit in corners, and may be about 12" deep at the point of the wedge. That 12" depth may consist of a 4" or 5" thick panel which faces out into the room, with two plywood side pieces, which form the triangular shape that fits into the corner. The entire corner bass trap is covered in fabric.
(Another type of bass trap doesn't involve a plywood panel. Instead, the entire 12" deep, wedge-shaped panel is composed of compressed rockwool. I understand that they can be as effective as the panel traps, which have air pockets behind them.)
Where a plywood panel is employed, the side pieces create a built-in air pocket behind the thick front-facing panel, so that the wedge-shaped bass trap can fit flush into the corner. As noted above, the thick bass traps, with air gaps behind them, can be at least somewhat effective down to about 60Hz. Bass traps are typically designed to also absorb mid and high-frequencies, in addition to the low-frequencies for which they were designed, unless otherwise specified. Those "broadband" bass traps can, therefore, also reduce ringing in mid-range and treble frequencies.
Sometimes, where bass traps can be partly effective with mid-bass frequencies (defined as about 50Hz to 120Hz), they can allow room correction to complete the job of smoothing-out peaks and valleys in the frequency response. (Room correction alone, is not always effective in doing that, nor can it meaningfully affect the overall decay rate in a room.) As with the higher frequencies, there is a good deal of personal preference involved in deciding how much bass trapping is required. Sometimes adding just a few bass traps enables a listener to achieve a clearer bass sound, without the sluggishness or boominess that he heard before.
There are a number of good sources for these more specialized acoustic panels, and they can also be DIYed. Most of the acoustic panel makers will offer free room treatment advice. I am showing a link to one well-known maker, but Ethan Winer's RealTraps, and ATS Acoustics are also good sources.
GIK Acoustics - Bass Traps with FlexRange Technology©
5. Location of the MLP:
With respect to the question about where to locate a main listening position (MLP), there are two factors to consider. One factor involves room modes, which only affect bass frequencies, and which mainly affect frequencies below about 250-300Hz. We will notice those room modes most with our subwoofers.
Typically, being at the exact center of a square or rectangular room is the worst possible listening position. 1/4 or 3/4 room length can often work well, but 3/8 or 5/8 of room length can sometimes be a little better. Those 1/8 of room length positions are theoretically "ideal" MLP positions, although measurable differences between those and 1/4 length positions can sometimes be minor.
If REW (an independent measuring program) or something such as the Audyssey App is used, it may be possible to differentiate between an MLP which has peaks at some frequencies, and one which has nulls (cancellation) at some frequencies. If it possible to differentiate between peaks and nulls, then a location with peaks will be easier to deal with than a listening position with nulls. For instance, an automated system of room EQ, such as Audyssey, can reduce peaks in frequency response. But, it can't do much with respect to deep nulls.
Again, experimentation is the key to discovering what works best, and as noted above, REW can be your friend in that process. Speaking generally, larger rooms tend to be a little more forgiving about room modes, and consequently about listening positions and subwoofer placement, than small rooms. A large room, in this context, might be about 4,000^3 or greater.
The second factor in determining a preferred listening position has nothing to do with bass frequencies. Instead, it involves early reflections of mid-range and high-frequencies. For example, in a large room, a listening position might be at 7/8 of room length and still be far enough away from the wall behind the MLP to avoid early reflections from that wall.
Ideally, you would want to be at least 1 1/2' to 2' from the rear wall, and 3' might be better. Otherwise, frequencies from about 1,000Hz and up may reflect from that wall in a way that garbles (distorts) the sound. That is because our brains can't easily separate sounds of equal volume, that arrive too close together. This was discussed in some detail in the previous subsection.
In a small room, however, it can be difficult to have a listening position which is 5/8 or 3/4 of room length (for bass frequencies) and still be 2-3' from the rear wall of the room, in order to prevent early reflections of higher frequencies. In that case, putting an absorbent panel on the wall behind the listening position can resolve the problem of early reflections.
Although most people on the subwoofer forum probably concentrate more on bass frequencies, it is important to understand that there can be two different objectives involved with respect to the location of the MLP. Sound quality for the higher frequencies is also an important objective.
Section I-C: Room EQ and Calibration Techniques
Systems of automated room EQ, such as Audyssey, measure the frequency response within a listening area, and then set filters (control points) at specific frequencies to equalize sound pressure level across the entire frequency range. Automated room EQ is generally believed to be helpful with bass frequencies (<500Hz) and especially helpful below the transition frequency in a room. The extent to which systems of room correction can correct higher frequency distortion in a room, caused by setup issues or by fundamental room issues, is a more controversial question.
In my opinion, this question can only be answered by individual listeners on a case-by-case basis. If the overall audio quality sounds better with Audyssey on, then the room correction is successful in that specific instance. And, if the overall sound quality sounds better without Audyssey, or some other form of room correction, then I see that as strictly a user preference issue. (Before writing-off a system of room EQ, such as Audyssey though, it is a good idea to experiment extensively with settings, such as DEQ. A number of user-controlled settings can influence the potential sound quality for a particular listener in a particular room. They are discussed throughout the Guide.)
* Note: Some systems of automated room EQ allow users to limit the use of room correction above a certain frequency, such as 500Hz. That is strictly a user preference issue. The Audyssey app, which is available with newer model Denon/Marantz AVR's/AVP's, has that feature, along with other user adjustability options. There is a thread devoted to that app, so I am not going to address it in the Guide. Instead, I am going to explain ways to maximize the sound quality of Audyssey calibrations in general. Readers who want to learn more about the use of the Audyssey app are encouraged to consult this thread:
MultEQ Editor: New App for Denon & Marantz AV...
There are some things that we can do initially, to enhance our chances of improving our audio quality, when we do a room correction calibration with a system such as Audyssey. First, we have to understand that there may be a limit to what Audyssey can do to correct pre-existing problems involving improper speaker setup, or too many early and late reflections. Better speaker setup, and even a modest use of room treatments, may augment what room correction can do to improve the sound quality.
Second, we need to understand that, in some cases, room correction can actually create distortion (or exacerbate it, if it already exists) if we don't observe good calibration technique when we EQ our rooms. So, this section of the Guide will offer some general tips which may assist users in getting better calibrations. Although the general focus will be on Audyssey, some of the tips may apply to other systems of automated EQ as well.
We need to realize that measurement microphones, such as the one employed by Audyssey, do not "hear" sounds in the same way that we do. To start with, the Audyssey measurement microphone is far more sensitive than our own hearing is. Small variations in volume at certain frequencies, which might completely escape our notice, will be picked up by the Audyssey microphone. And, Audyssey will try to "fix" them, even if they don't really need fixing, and, even if we would already have been able to hear those frequencies with subjectively good sound quality.
For instance, if speakers are toed-away from the MLP a little too much, Audyssey may detect that the high-frequencies are a little too low, in comparison to the frequencies which are not quite so directional, and Audyssey may boost those high-frequencies accordingly, causing some shrillness. The reverse can also be true, when some types of tweeters are pointed too directly at the MLP. For example, some manufacturers (or experienced product users) may recommend that specific front speakers not point directly at the MLP. That can especially be the case with some horn speakers, which have high sensitivity, and which may sound bright compared to other speakers.
In either case, Audyssey may exacerbate a barely noticeable (or completely unnoticeable) but pre-existing situation, by trying to do too much correction in a frequency range where a lot of correction may not be required for an individual to hear subjectively good sound quality. I think that Audyssey can actually be used as a kind of test instrument to help users discover how to point their speakers in just the right direction to maximize sound quality, although the resulting differences in SQ may be subtle.
I found that to be the case in my own room. A little less toe-in gave me more harshness in the upper frequencies, post-Audyssey. A little more toe-in, and Audyssey did less to affect those higher frequencies, resulting in better, more natural, sound quality. In my particular room, and with the specific setup and adjustments I use, Audyssey simply sounds better on that off for the full frequency range, although the most noticeable difference is still below 500Hz.
What I am saying here is that if we wish to fully benefit from automated room correction, and if we are really serious about the sound quality in our rooms, then it may be necessary to experiment a bit with speaker positioning, with appropriate softening influences, and then with subsequent Audyssey calibrations, in order to achieve the best overall sound quality that we can. It is a little bit of extra work initially, but the long-term results can be worthwhile.
We shouldn't just assume that the Audyssey microphone will hear exactly what we hear, because it won't. The sensitive omnidirectional Audyssey microphone will "hear" things that we won't. And, our brains will filter and influence what our ears do hear. We also shouldn't assume that it will EQ our systems in accordance with our personal listening preferences, if we don't experiment a bit with our speaker placements and our EQ technique when we are running our Audyssey calibrations. It's really just trial-and-error to find out what works best for a particular listener in a particular room.
[I should note here that my intent in writing this is not to make people obsessive about their speaker placements, or about any other factor being discussed. The real intent is just to inform people that factors such as speaker toe-in can potentially impact an Audyssey calibration. If the initial Audyssey calibration sounds good, then don't worry about it. If, over time, someone wonders whether any further improvements can be made, typically to the higher frequencies, then some additional experimentation might be helpful.]
A second example of the difference between what we hear, and what Audyssey hears, concerns the nature of the omnidirectional Audyssey microphone. The Audyssey microphone hears sounds equally in all directions, but we don't. The pinnae (flaps) in our ears funnel sounds into our ears from the front and from the sides. But, they partly block and deflect sounds coming from behind us. So, early reflections from a wall behind us are going to be noticed far more by the Audyssey microphone, than they are by our ears. And, in trying to over-correct those early reflections, Audyssey may actually contribute to the distortion we hear.
[As a general rule, we don't want Audyssey to become extremely busy in the high-frequency range. Reflections from hard surfaces, bouncing into the Audyssey microphone at close range, can cause Audyssey to set too many control points at high-frequencies, causing additional distortion which may sometimes be characterized as comb filtering. The more that we understand why Audyssey might be doing things to "correct" mid-range or high-frequencies, the more that we can enjoy the overall benefits of room correction for low-frequencies, without adversely affecting our higher-frequency sound quality.]
The nature of the omnidirectional microphone is why Audyssey advises keeping measurement microphones at least 18" away from a wall or other hard surface. Perhaps an even better example of the difference between the way we hear, and the way the microphone hears, involves chair or sofa backs. Most chairs or sofas in our HT's or mixed-use rooms have relatively smooth surfaces. Some of them have fairly firm leather surfaces.
Those smooth or firm surfaces reflect high-frequency sound waves directly into the Audyssey microphone, in a way that they never could if we were actually sitting there. And, the sounds from the back of a sofa would be sufficiently attenuated by our pinnae (ear flaps), and would reach our ears so simultaneously with the direct sound, that we would never hear them. But, the omnidirectional Audyssey microphone would hear them, and in trying to correct something that didn't need correcting, it could introduce comb-filtering (high-frequency distortion) into the sound.
One way to avoid that problem would be to keep the Audyssey microphone at least 18" away from a chair back. But then, we wouldn't be measuring where our ears are, and that could negatively impact our calibration. A better solution is simply to drape a fluffy blanket over our chair backs during calibration. That will enable us to get our microphone within about 4" or 5" of the chair back, and where our ears actually are as we listen. At the same time, that will prevent high-frequency sound waves from bouncing into the mic from very close range. And, Audyssey will leave those spurious high-frequency reflections alone, concentrating its EQ resources on broader areas of the frequency range. (Chris Kyriakakis, the creator of Audyssey, has endorsed this solution.)
Audyssey employs a system of fuzzy-logic weighting to average the results from either six or eight microphone positions (depending on the Audyssey version). In general, I believe that the more we can give Audyssey more consistent measurement results to work with, the more that we can achieve a smoother frequency response, and consequently improved sound quality. This is something that Chris Kyriakakis commented on in response to a question. He suggested that the more uniform the sound is within a measurement area, the more uniform the Audyssey EQ is likely to be. And, the more uniform the Audyssey EQ is, the more likely it will be to provide good sound quality to a larger listening area. (We should recognize that no system of automated room EQ is likely to be able to EQ an entire room. The smaller the listening area we are trying to EQ, the better our resulting sound quality is likely to be.)
** The examples above illustrate one aspect of microphone placement, within a more consistent measurement area, but general microphone placement is also a factor. We typically want to have our microphones at ear level, even if not all of our speakers are at that same height. Keeping the mic at roughly ear level seems to be consistently important. Some users, including myself, have achieved good results by taking just a couple of measurements 2" or 3" above ear level. We typically do not want to go behind a chair back with any of our measurement positions, unless we are deliberately trying to EQ for a second row of listening chairs. And, even then, it would be a good idea to experiment both ways--going behind the MLP, and not going behind it. It helps to recognize that Audyssey is simply trying to EQ a fairly uniform listening area, and not individual seats. Other than mic position 1, which is typically centered on the primary listening chair, the mic positions don't need to correspond at all to any actual seats.
Readers interested in why it might not typically be a good idea to measure behind the main listening position, if no one is actually sitting there (and often, even if someone is) may wish to read a post from later in the Guide thread. It adds a little more detail about the issue:
Guide to Subwoofer Calibration and Bass Preferences
It is a very good idea to use a boom microphone stand with an extendable arm for Audyssey calibrations. That allows the base of the stand to remain on the floor, while the swing-out arm allows the microphone to be positioned exactly where the listener wants it. If something such as a camera tripod is used, mic placement can be much more difficult, and the heavy mass of the tripod can interfere with the calibration due to secondary reflections from the tripod. If a tripod is placed in a chair or on a sofa, to facilitate mic placement, vibrations from the furniture can be passed up through the Audyssey microphone. Whether that would always make a significant difference in the calibration is somewhat debatable, although there have been some good before-and-after examples where it definitely did make a difference.
Another issue with camera tripods is the bulk of the tripod itself, positioned directly below the microphone. There can also be spurious reflections from the base of the tripod, which can interfere with the accuracy of the calibration. For both Audyssey and for REW, boom mic stands work much better. And, for the small cost involved, I think that it makes sense to use a much better and more stable stand than the cardboard one that Audyssey provides.
The type of stand I am recommending provides for much more exact mic placement, and better repeatability for calibrations. That repeatability is important, once someone has found a mic pattern that works well for his room and equipment. There are a number of different stands that can work, although I have heard that some of them which come with adapters included, are flimsy and don't stand-up well. The one that I am linking below is sturdy and durable, but it does require a separate adapter to hold the Audyssey microphone. I am also linking two different adapters. Either adapter can work.
Amazon.com: On-Stage MS7701B Tripod Microphone Boom Stand: Musical Instruments
Amazon.com: On Stage CM01 Video Camera/Digital Recorder Adapter: Musical Instruments
Amazon.com: On-Stage MY200 Plastic Clothespin-Style Microphone Clip: Musical Instruments
*** As a general rule, it is a good idea to measure smaller areas, as opposed to larger areas, for the reasons cited above. We want our measurement area to be large enough to accurately represent the binaural nature of our hearing. (For bass frequencies which are low enough to be non-directional, hearing them with just one ear may be sufficient.) For all frequencies though, we at least want to measure in a fairly large circle forward of and out to the sides of our heads. Fairly large, in this case, could be a circle about 18" to 24" in diameter.
But, we may not want to measure such a large area that we present Audyssey's fuzzy-logic weighting system with too much anomalous information. The more uniform the frequency response is, within a measurement area, the better the resulting calibration is likely to be. Patterns that vary in size from as small as about 6" to 12" out from the MLP (mic position 1), to as large as about 24" to the side and forward are typically used. I would not generally recommend going further out to the sides, or forward more than about 24" from mic position 1.
It is interesting to note that, in the last couple of years, Audyssey has revised it's owner instruction manuals to recommend a smaller microphone pattern than they used to recommend. They used to recommend 3' to 4' out from the MLP. I believe that they now recommend about 2' or less. Their revised recommendations seem to parallel the experience of many Audyssey users, who discovered that smaller microphone patterns often resulted in better sound quality, over a wider area, than large mic patterns did. That is consistent with my own experience, and with that of a number of others on the Audyssey thread.
But, I suspect that finding an optimum microphone pattern is at least somewhat room and system dependent, so I suggest that interested users experiment in an effort to discover the specific microphone pattern which produces the best sound quality in their rooms. Once they find a mic pattern that they really like, I recommend that they write it down, or draw it, so that they can return to it for future calibrations. Sometimes, fairly subtle differences in microphone placement can yield significant differences in the resulting sound quality.
**** For people who are looking for some preliminary guidance in selecting microphone positions, the following visual aid is offered. This roughly 2' by 2' pattern is one that a number of people have successfully used. But, it is only shown as a starting point and not as a specific recommendation. People still need to experiment to discover what pattern works best in their particular circumstances.
In this pattern, mic position 1 is about 4" or 5" in front of a blanket covered chair (or couch), the center of which is the MLP. Remember that MLP stands for main listening position, and mic position 1 is always, by definition, the MLP. The MLP can be the center of a couch, or the center of a chair, depending on the specific room. For purposes of the illustrated diagram, mic position 1 is right between your eyes (and ears) and about 4" or 5" way from the blanket covered surface of a chair or couch. The mic is at a height which approximately corresponds to the center of your ear canal, as you would sit when listening to music or watching movie. That is what we mean by ear height.
It may be important to to note here that mic position 1 is used to set volume levels and timing (distance) for all of the channels. In order to accomplish that, Audyssey only uses a portion of the full bandwidth sweeps in mic position 1. It uses the 30Hz to 70Hz bandwidth to set levels for the subwoofers in mic position 1, and it uses the 500Hz to 2,000Hz bandwidth for the other channels. Full bandwidth sweeps of 10Hz to 22KHz are employed for all of the mic positions in order to set EQ filters. It is important to keep all of the mic positions fairly close together in order to insure that Audyssey's system of fuzzy logic weighting is presented with somewhat uniform measurement information.
Positions 2 and 3 are out to each side of 1 by about 10" to 12". Positions 4 and 5 are straight out in front of 2 and 3 by about 20" or 24". Number 6 is in a straight line out from 1, but this time only about 14" to 18" away. All six of those mic positions are right at ear height. Positions 7 and 8 are in fairly close to the chair back--perhaps only about 6" away from the blanket and about 6" out to the side of mic 1. (That clusters some mic positions very near the head, and where the ears on each side of our heads are located.) Both of the last two positions can be raised up by 2" or 3" above ear level. In this particular mic arrangement, none of the mic positions go behind the chair.
2--------1---------3
-----7--------8-----
---------6-----------
4-------------------5
The specific order of the mic positions is not important, so after mic position 1 (which is always the MLP) the numbers assigned are arbitrary. Users can follow the diagrammed positions in whatever numbering sequence works best for them. It is only important to keep the mic level (so that it points upward) and close to ear height for at least about the first six positions. People who have a version of Audyssey which only uses six mic positions might wish to eliminate 7 and 8 from the diagram shown, or they could experiment with an even more compact configuration for their six. Experimentation is the key to finding a result which pleases the individual user.
Section II: Audio System Calibration and Subwoofer Levels
This section explains how Audyssey calibrates and EQ's our audio systems, and offers some advice for the best methods to boost our subwoofers. It also explains what Dolby/THX Reference is and how that standard specifically relates to the subwoofer boosts we may prefer.
The subsections in Section II are as follows:
II-A: Audyssey Calibration and Dolby Reference
II-B: Why We Add Bass After Calibrations
II-C: Where and How to Add Bass
II-D: Master Volume Levels and Sub Boosts
II-E: Gain Settings and Maximum Output
Section II-A: Audyssey Calibration And Dolby Reference
Audyssey Calibration:
Audyssey is a room correction software program designed to reduce room/speaker interactions which may adversely affect audio quality. Once we put a speaker or a subwoofer inside a given room, it's native frequency response changes, depending on a number of factors, including its specific placement in the room. Audyssey attempts to remove undesirable room influences by setting control points to even-out the frequency response, so that we don't have large dips or peaks in sound pressure level (SPL) at certain frequencies. Audyssey's goal is to make all frequencies approximately +/- 3dB from a standardized calibration SPL of 75dB.
There are multiple versions of Audyssey which have been introduced over the years. The newest and best version of Audyssey room correction (not counting the Audyssey App which simply permits more user control of what is being EQed) is Audyssey XT-32. The following table shows the various versions of Audyssey.
The versions are shown relative to 'X' filters in 2EQ, where 'X' represents 8 filters in that very early version. (We frequently speak of the numbers of filters used in room EQ. But in reality, there is only one filter per channel, and each filter has 'X' number of control points that it can employ.) The latest version of Audyssey, XT-32, has 4,096 control points available per channel. That is 8 x 512.
Audyssey, in all current versions performs its room correction by sending 75dB test tones to each of the channels in an audio system, and by then measuring the frequency response for each channel, from 10Hz to 22KHz. Once the measurements have been completed, Audyssey calibrates the results, using a system of fuzzy-weighted logic. It then sets control points at individual frequencies, or groups of frequencies, in order to correct for peaks and valleys in the sound. Once the control points have been implemented, the room EQ is complete.
* When Audyssey, and other systems of automated calibration, perform a system calibration they ignore all prior settings, such as the master volume level, trim levels, distances, and crossovers. Once the calibration is complete, the master volume level will typically return automatically to the volume that the AVR was on prior to the calibration.
It is important to note that all measurement microphones, including the Audyssey microphones, have an inherent error factor of anywhere from +/- 1.5dB to +/ -3dB. (The Audyssey microphones have an error factor of +/- 3dB.) That includes very expensive calibrated microphones used to measure SPL or frequency response, such as the UMIK-1.
If you measure the Audyssey test tones, or your post-Audyssey SPL, with a calibrated microphone, you may find that your SPL measures two to three decibels above or below 75dB. I think it would be fairly rare for two different microphones to measure exactly the same, in any case. The important thing is that all of the channels in your system are playing the same SPL, as measured at the MLP, and Audyssey will be quite accurate in that respect.
(It is not a good idea to double-check the post-Audyssey volume settings for the various channels, using your AVR's internal test tones. Instead, it's much more accurate to use external test tones, through a test disc. The internal test tones in Denon/Marantz AVR's bypass the filters that Audyssey set for the various channels. The test tones are in a completely different software program than the EQ program. It would not be uncommon for there to be a difference of a decibel or two between a calibrated speaker level and the uncalibrated one found in the test tones. The test tones are intended only for manual adjustments in trim levels.)
It is important to understand that two separate actions are performed during a calibration process. First, the system will be calibrated to Dolby Reference using 75dB test tones. All channels (including subwoofers) are set to have the same sound pressure levels, as measured at the main listening position (MLP), and all sounds are set to arrive at the same time, via the distance settings. Preliminary crossovers will also be set during this process. (As noted in the Cliff Notes, the preliminary crossover set by an AVR is not a recommendation. It is actually just an observation, as the AVR uses it's own default programming to make the preliminary crossover settings.)
The second thing that will occur, during the calibration, is the actual room EQ process. In that process, Audyssey will measure the frequency response from various microphone positions, and will use a system of fuzzy logic weighting to set filters for all of the channels, in order to remove some of the peaks and valleys in the frequency response that inevitably occur whenever transducers (speakers or subwoofers) are played inside a typical home theater or mixed-use room.
It is important to distinguish between the two processes. The initial calibration process insures that equal volume levels, from all of the channels, will arrive at the MLP at the same time, and it calibrates each of the channel volumes to a Dolby/THX Reference standard. The second part of the calibration process is the room EQ process, which sets filters for all of the channels, in an effort to improve the overall sound quality in the room.
The room EQ software program, and the filters (control points) it applies to an audio system are independent of the various AVR settings. So, once the calibration is complete, changing any settings will not change the room EQ that Audyssey applied to an audio system. This is a recurring question, so I have emphasized it here. AVR setting changes do not affect the room EQ that Audyssey applies. And, Audyssey can be turned off, and then turned back on again, as often as we like. Once Audyssey is turned back on, the same room EQ will be applied. (Using the Pure Audio mode will disable Audyssey. When Audyssey is disabled, features such as DEQ and Dynamic Volume are not available.) If significant changes are made to the room, however, or to any of the speakers or subwoofers (new locations, for instance) a new calibration should be performed.
One of Audyssey's goals, in any Audyssey version, is to set the volume levels of all channels in a system, including subs, to 75dB, as measured at the MLP, by the calibrated Audyssey microphone. The MLP is microphone position 1, by definition, wherever the user chooses to place the microphone. And, that point in space is where Audyssey will set timing (distance) and trim levels (volume levels), for all of the channels, to coincide.
The .1 in a 2.1, or 5.1 (or larger) audio system is the LFE (low frequency effects) channel, and that .1 designation has nothing to do with the number of subwoofers in a system. The .1 designation was originally selected because the LFE channel only plays a fraction of the total frequency range of an audio system. (The LFE channel is explained in a little more detail in Section III.) Where there are subwoofers configured in an audio system, Audyssey will measure and calibrate all of the subs in a multi-sub system together, so that their combined SPL is 75dB. Whether there is one subwoofer, or there are many subwoofers in an audio system, the combined sound of the subwoofers as a whole will be set to 75dB.
[The difference between subwoofers and the .1 LFE channel may be a little confusing at times to all of us, so I decided to add some clarification to this section. Most people reading the Guide will have subwoofers connected to sub outs in their AVR's or AVP's. Those AVR's and AVP's are designed to play Dolby 5.1 programming, which contains the .1 LFE channel, briefly described above. But, it is possible to connect subwoofers to some stereo amplifiers. We can also listen to 2-channel content, or watch older movies (made prior to the development of Dolby 5.1, which occurred in 1992) on our HT systems. However, the audio system is only playing LFE (.1) content, which is recorded 10dB louder than the regular channels, when 5.1 or higher material is being played on an AVR or AVP.
For any other program material, the subwoofers only play content contained in the regular channels (such as 2-channel content) even if the content is upmixed using a surround mode such as Dolby Pro-Logic (PLII). When subwoofers are employed for non-native 5.1 material, they simply enhance the bass in the regular channels by playing frequencies below the crossovers assigned to those speakers. It is important to understand the distinction I am making, because all of us tend to equate subwoofers with the .1 LFE channel, and they are two different things.
The .1 channel is for low-frequency effects content (special bass effects), recorded 10dB louder than other content in a 5.1 soundtrack. Subwoofers are transducers specifically designed to play bass frequencies, below crossover points, in the regular channels. And, in addition to that function, they also play .1 LFE content, whenever a 5.1 program is played through an AVR or AVP. It is worth mentioning that the .1 in 5.1 or 7.1 doesn't refer to the number of subwoofers. The .1 designation was used because the LFE channel was assumed to be playing about 1/10 of the total audio content in a movie.]
The trim levels and distances (timing) for all of the channels (including the subwoofers) will be set at mic position 1, from the initial test tones at that first mic position. Audyssey will disregard any previous settings and set levels, distances, and crossovers from scratch, whenever a new calibration is run. As noted, trim levels and distances for all of the channels are set based on microphone position 1. Crossovers are set after all of the test tones are completed, based on a fuzzy-weighted average of all of the microphone positions in a calibration.
The sweeps which Audyssey uses during its calibration process cover a frequency range from 10Hz to 22KHz. (A recent Audyssey zendesk answer stated that the sweeps go up to 24KHz. It may be that Audyssey's sweep range has changed in recent years, but if so, it won't make any difference whatsoever for the frequencies which humans can actually hear. So, I will keep using the range of 10Hz to 22KHz for the Guide.)
All trim levels and distances are set before Audyssey adds control points to the channels. And, all internal test tones, which govern those trim levels, remain independent of the EQ which Audyssey performs. The Audyssey EQ process occurs automatically once we tell Audyssey to "Calibrate". Audyssey takes the data from all of the mic positions and applies its fuzzy logic weighting, as previously described.
[The sweeps that Audyssey uses, during that calibration process, provide an interesting illustration of the way that our hearing works. The sweeps (broadband test tones) used for the regular channels sound much louder than the sweeps used for the subwoofers. But, all of the sweeps are playing at the same 75dB volume.
All of the sweeps, for all of the channels, are playing the same frequencies, but we aren't hearing them the same way. At a volume level of 75dB, the subwoofers can't go very far up into our normal hearing range with audible sounds, and that will make them seem much softer compared to our speakers which are playing sounds in our normal hearing range.
We may also hear a difference in the test tones, with larger speakers, which may sound lower in tone than smaller speakers do. That would be a product of the low-frequency capabilities of the speakers. And, we may also hear some difference in loudness between more distant speakers and closer speakers. Inside a room, bass frequencies lose about -3dB of volume for every doubling of distance. Frequencies above about 300Hz lose -6dB for every doubling of distance, due to the fact that the higher frequencies are not benefiting from boundary gain (from the walls, floor, and ceiling) in the way that bass frequencies are.
But, the biggest difference in both tone and loudness will be between the sweeps for the regular channels, and the sweeps for the subwoofers. The difference in the relative loudness between the sweeps for the regular channels, and the sweeps for the subwoofers, is the difference between the way we hear frequencies in our normal hearing range, and the way that we hear low-frequencies. Frequencies in our normal hearing range of about 500Hz to 5,000Hz sound much louder than the frequencies played by our subwoofers.
The test tone sweeps give us a good example for why most of us need to add more bass, once all of our channels are level-matched. That is especially true as our overall listening levels drop, because low-frequencies drop-away faster, relative to those in our normal hearing range.]
As noted above, the first microphone position is also used to set distances for all of the channels. Audyssey sets distances for the channels based on the time it takes for the sound to arrive at mic position 1. Subwoofers have their own internal amplifiers, and their own internal processing, which typically delays the timing of the sound arriving at the MLP. Audyssey compensates for that delay by setting subwoofer distances as longer than the physical distance from mic position 1. Setting a greater distance for a channel causes the AVR's internal programming to speed-up the arrival time of the sound (and vice-versa). In the subwoofer's case, speeding up the signal, with respect to the sound from the other channels, allows the sounds from all of the channels to arrive simultaneously at the the MLP.
Dolby/THX Reference:
When Audyssey finishes, all channels in the system (including all of the subwoofers' combined SPL) will play at the same volume, at the MLP, as determined by the calibrated Audyssey microphone. And, when all channels in a system are playing at the same volume, the sound at the MLP will be approximately in balance with what the film mixer intended whenever a movie is played at "Reference" volume, which is 0.0 master volume. However, when the listening level is lower than about -5 MV, most people will not hear bass frequencies quite as well as other frequencies, or quite the way that a film mixer intended for them to be heard in equilibrium with the other frequencies in the film.
The Dolby, or THX Reference standard is intended to provide some degree of uniformity in the maximum volume levels of movies, and is intended to provide a way for commercial cinemas and home theaters to make sure that their audio systems correspond to what film mixers intended for them to hear. The Reference level is capped at a peak volume of 105dB for the regular channels, and 115dB for the LFE (low frequency effects) channel. Those were considered maximum safe listening levels for movies. Most of those peak volume levels of 105dB for the regular channels and 115dB for the LFE channel would be for very short durations.
The LFE channel enables additional low-frequency effects, below about 120Hz, to be mixed into a 5.1 movie (or music) track. A system that is calibrated to Reference will, if it has the capability to do so, automatically play those 105dB and 115dB peaks when the master volume control is set at 0.0. It is important to note that not every system is capable of playing those 105/115dB levels, nor would those volume levels be found in every movie. (The .1 LFE channel is explained in some detail in Section III-B.)
As stated, Dolby/THX Reference is just a guideline that provides a degree of standardization for the way that movies are recorded, and for the way that commercial cinemas and home theaters are calibrated. But, all the Reference standard really does is to establish maximum volume levels for the regular channels (105dB) and for the LFE channel (115dB). There is no specific uniformity with respect to the average volume level of movies, and there is no specific uniformity with respect to how frequent, or how sustained, crescendos of up to 105/115dB will be in a particular movie. This is an important point to understand!
Some movies may be much louder than other movies, both in terms of the average volume level of the film, and in terms of whether the film actually hits crescendos at the upper limits of the Reference standard. Some movies may never hit the upper limits of the standards, and some movies may hit those upper limits again and again. That is entirely at the discretion of the director and the film mixer. That is easy to understand if we compare blockbusters to light romantic comedies. But, even among different blockbusters, or among action films in general, there may be volume differences. And, there may be profound differences in how low in frequency the movies goes. So, overall volume, peak volume, and very-low-bass volume, can all be variables among different movies.
Once an audio system has been calibrated to Reference, how much or how little of the total SPL capability of an audio system is actually employed, is entirely a matter of personal preference. Some people prefer to listen at much louder levels than others do. It is important to understand that Audyssey will set all of the channels in a system to play at equal volumes at the MLP. And, when the master volume is at 0.0, the audio system will be playing Reference volumes, if it is capable of reaching those sound pressure levels. But, it is up to the individual user to decide how loudly he wants his audio system to play, via the master volume control.
That same principle also applies to our subwoofers. Sometimes, people will add a second subwoofer (or a larger subwoofer) to a system, and then question why the bass doesn't sound any louder after the system is recalibrated with the greater amount of subwoofer headroom included. Having multiple subwoofers increases our potential bass headroom, so that we can hit higher bass SPL's. But as noted earlier, the combined SPL of our subs will always still be calibrated to 75dB, so that the combined volume of the subwoofers will be equal to the volume levels of the other channels in our system. In order to actually use any additional bass headroom we may have available, we either have to play our audio system at higher listening levels than we were before, or we have to increase the volume level of our subwoofer(s). Or, we can do both.
At one time, test tones were employed at the "nominal" average Reference volume of 85dB. (I use the term nominal average volume, since there really isn't an actual average volume level for movies. Film mixers just record movies at the volume level that seems appropriate to them, although some TV channels do impose strict standards on average and peak volume levels.) The 85dB number was selected when the Dolby/THX Reference standards were developed, to represent a hypothetical average, with 20dB of headroom above that for the regular channels, and 30dB above that for the LFE channel. That same 85dB number was chosen as a uniform calibration number. And audio systems were calibrated to Reference at 0.0 master volume with 85dB test tones.
However, the original 85db test tones which were used were uncomfortably loud for most people, so most systems of automated or manual calibration, including Audyssey, converted to a less uncomfortable test tone of 75db. And, all channels are calibrated equally to that 75db level. Our AVR's then do an internal recalculation to add 10dB to the regular channels, so that a master volume of 0.0 will equal approximately 85dB. And, once that internal recalculation takes place, the audio system is now calibrated to Reference (allowing for 105dB max for the regular channels, and 115dB max for the LFE channel) at a master volume of 0.0.
Once an audio system has been calibrated to Reference, whether that particular audio system will subsequently be able to play 105dB peaks in the regular channels and 115dB peaks in the LFE channel is entirely a function of the capabilities of the individual audio system in that particular room. And, as mentioned earlier, whether an individual user will ever decide to play his audio system at Reference levels, even if the system is capable of doing it, is entirely an individual choice. In fact, most people don't ever play their audio systems that loud.
[Based on anecdotal information, the average HT listening level is probably in a range from about -20 to -10 MV, with a master volume of -10 sounding twice as loud as a master volume of -20. (Each additional 10dB of SPL equals a doubling in perceived loudness.) Many people will be either below or above that average range, but most of us probably fall within it.
As noted in Section I, a number of audio experts have determined that rooms under ~20,000^3, which would include almost all home theaters, sound anywhere from +5 to +9dB louder at the same SPL than would be the case in a commercial cinema. Commercial theaters are much larger than home theaters, and higher sound levels don't sound quite as intense in those large spaces. That probably helps to account for the fact that relatively few people, even in treated rooms, listen at 0.0 MV.
Regardless of preferred listening levels, however, it is important to have all of the channels in an HT system playing at the same volume at the MLP, so that sounds from all of the speakers (and subwoofers) will be in proper balance. But, as a practical matter, starting with all of the channels (including the combined sound of all the subwoofers) playing at the same volume is probably also the only way to set the audio system to Audyssey Flat. The intent of the Flat response curve is to have every frequency from down to as low as 10Hz, and as high as about 22KHz, play +/- 3dB. Audyssey attempts to accomplish this by setting control points which add boosts to some frequencies, and cuts to other frequencies, within every channel in an audio system.
Speakers and subwoofers are typically designed to play a reasonably flat frequency response, so that some frequencies don't stand out in comparison to others. But, once those speakers, and especially subwoofers are placed in a room, the room itself will affect the frequency response, causing peaks at some frequencies and dips at others. Proper speaker and subwoofer placement within a room will help, but bass frequencies in particular often need help from some form of room EQ to even-out the frequency response.
Audyssey attempts to provide that help by setting control points within every channel. When an audio system has been EQed to a relatively flat frequency response, the room is at least partly taken out of the equation, allowing the speakers and subwoofers to play a naturally flatter frequency response. This is generally believed to be particularly helpful for bass frequencies (from about 500Hz down), and may also be helpful for higher frequencies, depending on the particular room and listener preferences.
The Audyssey Reference curve changes Flat by creating a slightly downward curve to the high-frequency response. It is the default setting after an Audyssey calibration. The Reference curve slightly rolls-off the treble frequencies above 4,000Hz (by about -2dB) and it adds more roll-off (about -6dB total) above 10KHz. It also adds mid-range compensation (a -3dB dip centered between 2,000Hz and 3,000Hz). Many people prefer those high-frequency roll-offs, and it is strictly a YMMV issue as to which of the two Audyssey settings we use. But, to create either the Flat curve, or the Reference curve, Audyssey needs to start with all channels and frequencies playing at the same volume at the MLP.
[Again, it should be noted that a similar methodology, of setting all channels, including the combined sound of the subwoofers, to the same volume, is used by other systems of automated or manual calibration, whether any room EQ is being attempted or not.]
** Occasionally, someone may say that he prefers to hear the "natural" sound of his speakers, or subwoofers, without any room correction at all. As far as I am concerned that is an entirely personal decision. There isn't any right or wrong way to listen to audio, in my opinion. But, I have never really believed that, given a good calibration, Audyssey is likely to significantly change the "natural" sound of speakers. Horn speakers still sound like horn speakers, and electrostatic panels still have their own characteristic sound. What Audyssey may do, however, is to change the way that speakers and subwoofers interact with a room. After all, changing room/transducer interaction is the purpose of implementing room correction. Whether that change is positive or negative is up to the individual to determine.
What any system of room correction attempts to do is to change the interaction of the speakers with the room, in order to allow the speakers (and subwoofers) to play with less distortion, caused by room-induced peaks and dips in the frequency response. As noted below, those peaks and dips are especially prevalent in bass frequencies. Again, whether Audyssey or any other system of automated room EQ is successful in creating a smoother frequency response, and whether attempting to do that results in improved sound quality, is strictly up to the individual listener to decide.
And, that may depend on variables which include the specific room, and room furnishings and treatments; the speakers/subwoofers, and their specific positioning; the relative care taken during the calibration process; the AVR settings employed; and the personal preferences of the particular individual.
But, in order to fully experience Audyssey, I would encourage users to take pains in system setup and in their Audyssey calibrations. As noted in Section I, differences in system setup can yield different results, as can differences in microphone placement. Some calibrations may sound and/or measure better than others, and once a good calibration routine is developed, it is a good idea to keep a record of the mic placements that produced the most positive results. I would also encourage users to experiment thoroughly with various settings, such as the Audyssey Reference curve versus Audyssey Flat; with DEQ on and off; and of course with bass boosts. Individual setting changes, such as those, can sometimes significantly change the nature of the sound.
Section II-B: Why We Add Bass After Calibrations
The room strongly influences bass response, due to the action of room modes, causing peaks and dips at various frequencies. That is why Audyssey can be so helpful in EQing subwoofers. Audyssey can implement boosts up to +9dB and cuts up to -20dB, at selected frequencies, in all of the channels, in order to achieve a flatter frequency response. When Audyssey is successful at flattening out most of those bass dips and peaks (at least to some extent) the result may be a smoother, clearer, and more uniform sound.
That less distorted and less boomy sound, without some frequencies peaking at much louder SPL's than other frequencies, may give the impression that there is less overall bass playing. And, there may actually be less bass playing, at some frequencies, if a particularly noticeable frequency (say at 50 or 60Hz) were peaking quite loudly prior to EQ. Just hearing all of the bass frequencies, in better equilibrium with each other, may contribute to the initial impression that there is less bass.
But, setting aside whatever impressions of lower bass volumes we may have, when we hear less distorted bass in our rooms, there is more to it than just hearing a smoother frequency response. Most people don't listen at Reference Volumes (0.0 MV) which is where the low-frequency content in 5.1 movies was mixed to be in correct balance with other frequencies. Once the volume level of a movie is reduced below Reference, in a typical home theater, those low-bass frequencies will typically be harder to hear, in relation to the frequencies where our hearing is stronger.
That is because, as the volume level drops, it will appear to drop faster for frequencies outside our normal hearing range. And, that particularly includes low-bass frequencies, which carry much of the special audio effects in movies. So, in a properly calibrated audio system, listening at anything less than very loud volume levels, it is pretty normal to perceive the bass as sounding too soft.
[It is worth noting that people sometimes attempt to double-check their subwoofer volumes, after an Audyssey calibration, with an SPL meter. Even if the SPL meter is correctly set to C-weighted Slow, however, it may not be able to accurately record sound pressure levels below about 40Hz, where Audyssey is measuring and correcting SPL. Unless the SPL meter is a slightly more expensive one, which is calibrated, it may not yield correct readings for subwoofer frequencies. That is particularly true for smartphone apps, and for the inexpensive meters sold at places like Radio Shack. They can be off by as much as -10dB or more at low-frequencies.]
After the level-matching process from mic position 1 is complete, the low frequencies (which, as noted, are harder for us to hear) are playing at the same volume as all of the other frequencies. This phenomenon of lower frequencies being harder to hear than higher ones (except for very high-frequencies) is well documented. Our hearing is strongest from about 500Hz to about 5,000Hz. So, frequencies played by our subwoofers may require more volume than the frequencies played by our regular channels. Some additional explanation of this is included in the section on DEQ, and in the Addendum on the thread history. A more complete discussion of the Equal Loudness Contours, which define our perception of loudness at different frequencies, has also been included in Section VII-C.
[For the sake of this discussion of bass boosts, it should be noted that the Equal Loudness Contours, which are a slight modification of the original Fletcher Munson Curves, are based on averages in normal, healthy human hearing. It may be assumed that, as with other human attributes, hearing generally follows the shape of a bell curve, with some individuals hearing bass frequencies somewhat better, and others hearing those same frequencies somewhat worse than the average. And, our hearing may (will) change somewhat as we age. Therefore, a given individual may be able to hear lower frequencies relatively better, or relatively worse than would be predicted by reference to the Equal Loudness Contours. Consequently, the final bass levels that we pick may be partly a function of our individual hearing capabilities, and also partly a function of our specific psycho-acoustic preferences.]
Although a lot of the discussion so far has focused on movies and on Reference levels, our desire for stronger bass may not be limited to 5.1 movies or TV shows at below Reference levels. Even if we are watching a 5.1 movie at Reference levels, some of us might still prefer to have more bass than what Audyssey provides with a flat bass frequency response. DEQ won't add any bass at all at a listening level of 0.0. The decision of how much bass we want to hear (at any listening level) is an entirely personal one, which may depend on a number of factors.
We may also prefer to apply bass boosts to music. And, we may wish to add sub boosts for some of the same reasons that we would add them for movies. Even in the absence of special effects in movies, we may not hear low-frequencies in music as well as we hear those in our optimal hearing range, or we might just prefer more bass, period. That would be especially true at lower listening levels. And, some music and some TV shows may have less low-bass in the soundtrack to begin with. For instance, some older music may have very little content below 60 or 80Hz, due to the nature of the recording process, and we might be used to hearing more low-bass in our HT systems than that. So, some people might wish to boost the subs just to hear what sounds like a more appropriate amount of bass.
[I think it is important to emphasize that the degree of sub boost selected for any program content is entirely a matter of personal preference, as is the overall listening level. And, individual preferences may change, as we go from one source to another, from one song or movie to another, or depending on our moods, from one day to another.]
Audyssey's DynamicEQ (commonly abbreviated as DEQ) is a separate software program which boosts the low-frequencies in all of the channels, including the .1 subwoofer channel. It also slightly boosts the high-frequencies in the regular channels. It is engaged by default whenever an Audyssey calibration is run. The boosts that DEQ adds are intended to, at least partly, compensate for the inherent difficulty in hearing lower frequencies (and to a lesser extent, high-frequencies) at below Reference volume levels. DEQ is explained in detail in Section V.
How much boost DEQ adds varies depending on the MV selected, with more boost added as listening levels reduced below Reference (0.0 MV) at a rate of about +2.2dB per -5dB below Reference. So, at -15 MV, for instance, DEQ would add a little over 6dB (6.6dB) of bass boost to all of the channels, including the sub channel.
Whether DEQ fully restores bass equilibrium to movie soundtracks is an interesting question. Most people seem to prefer more bass boost than DEQ provides, and typically add an independent sub boost, even with DEQ on. With DEQ engaged, the typical sub boost appears to average about +3dB to +6dB. With DEQ off, sub boosts are typically much larger. Additional information, regarding DEQ, may be found in a later section. But, with or without DEQ, the question of how and where to add a sub boost is important for most people.
Section II-C: Where And How To Add Bass
Most modern commercial subwoofers have a gain (sometimes labeled "volume") control. That gain control helps to determine how much power will go from the sub amp to the driver (woofer). The way it works is that voltage goes from the AVR's sub out amplifier, to the subwoofer amplifier, which then amplifies that signal according to where the gain level of the subwoofer is set.
During the Audyssey calibration, the initial setting of that gain control will determine where Audyssey sets the trim level for the sub(s). So, if the initial gain control is high, Audyssey will set a low trim setting in the AVR (such as -9) in order to insure that the sub is playing 75dB at the MLP, just as all the other channels are. If the gain control setting is low, Audyssey will set a high trim level in the AVR (such as -3.0, or 0.0, or even +3.0) to insure that all of the subwoofers in a system are playing at that same 75dB level. A simple way to think of what happens during the initial calibration is: high gain level = lower trim level; low gain level = higher trim level.
[To emphasize this point more clearly, during the automated calibration process the subwoofer gain level and the AVR trim level are inversely proportional. For every decibel that we increase the gain level on our subs, the AVR will subtract one decibel from our AVR trim level. So, if we start with our subwoofer gain level at X, where X = 75dB, the AVR will set our subwoofer trim level at Y. If we set our subwoofer gain level at X + 5dB (80dB) our AVR will set our trim level at Y - 5.
That is because it is the AVR's job to make all of the channels play that same 75dB volume. Otherwise, our HT system can't actually be calibrated to the Dolby/THX Reference level that was our original target. The same thing would be true, in reverse, if we set our subwoofer gain level below 75dB, to 70dB, for instance. Our AVR would raise the AVR trim level by +5dB to compensate for the low volume of the subwoofer amplifier. Doing that would insure that all of the channels were calibrated to play that same 75dB at the MLP.]
So, that explains the relationship between the gain control and the AVR trim level during the initial calibration. But, what about after the calibration? If we want to add a subwoofer boost after Audyssey has set all of our channels to play at exactly the same volume level, how should we do it? should we use the gain control, or should we use the AVR trim control? We can actually use either one, but the decision of which one to use in a particular situation is just a little complicated.
The first thing to understand is that it is often desirable to make the subwoofer amplifier amplify the signal which is sent to the driver, rather than trying to have too much voltage coming directly from the AVR amp, because the subwoofer amplifier is much more robust and powerful than the amp in the AVR. This can be an extremely important point, because using the subwoofer amp for substantial volume levels may help to prevent clipping of the pre-out signal coming from the AVR. Clipping is a form of distortion, due to an alteration in the waveform. It can be audible in some cases, and if prolonged, can lead to overheating the voice coil in the woofer. When a waveform is clipped, the round top of the wave is squared-off---clipped-off.
That can happen when a subwoofer attempts to play low-frequencies with too much voltage. Simply raising the trim control in the AVR may not result in sufficient clean voltage going from the AVR pre-out to the subwoofer, if the subwoofer gain is set too low. Depending on the AVR, the pre-out signal may be slightly higher or lower, but in all cases that pre-out signal will not be as robust as what the subwoofer amplifier can produce. So, turning-up the trim may not actually result in more undistorted voltage going to the subwoofer. Instead, the AVR may send out a clipped signal at higher trim levels. That's why you typically want to avoid having too much voltage coming from the AVR, and insufficient amplification of the signal, occurring in the subwoofer amplifier.
It is necessary to have enough voltage coming from the AVR amp to turn-on the subwoofer from its Auto-On mode, and that can vary a little bit among some older AVR's. The -5 AVR trim number is sort of an arbitrary number, but several subwoofer makers suggest using that as about the max AVR trim setting for subwoofers. Holding our AVR trim at about -5, or lower, forces us to use the gain controls on the subs to add any really substantial subwoofer boosts.
After running Audyssey, simply making any adjustments in sub boost using the gain control on the sub(s) would insure that the sub amp is being used. So, that would be a perfectly good way to add sub boost. And, as noted a little further down, using a higher gain control may enable some subwoofers to achieve higher SPL's than they can with low gain levels and high trim levels.
But, most people find it more convenient to make adjustments using the AVR trim controls, with a remote control. And, in that case, it is desirable to start with a high sub gain level, and a low AVR trim level. Remember that a high gain level = a low AVR trim level. So, we would need to take certain steps during the calibration process if we wanted to have a low AVR trim level--let's say in about the -10 range. And then, we could adjust the trim level upward after the calibration. Using the trim settings in the AVR to make sub volume adjustments, after running Audyssey, allows the user to make convenient and fairly exact (.5dB increments) adjustments to subwoofer volume, by using the AVR remote.
Typically, in order to achieve a low AVR trim level, though, it will be necessary to start with a measured sub SPL of higher than 75dB. An SPL level of about 78dB to 80dB may be required. That would be in the red zone for Denon/Marantz units during the subwoofer level-matching process. Audyssey is specifically trying to set the sub(s) SPL to 75dB. That is in the green zone on Denon/Marantz. However, to get a strongly negative trim level, a higher than 75dB level will be required, and that will be in the red zone. The specific SPL used is not as important as the resulting low AVR trim level.
It should be noted that there is no harm in telling Audyssey to proceed with the calibration, even though the subwoofer is not in the green zone. That notification of red zone just gives owners the opportunity to adjust the gain on the externally-powered subwoofer to the same 75dB volume level which Audyssey is trying to achieve for all of the channels. But, if the owner chooses to proceed with the calibration anyway, Audyssey will simply calibrate the subwoofer(s) with a low trim level, and that is typically exactly what we want it to do. That way, we can start with a low AVR trim setting, and add some subwoofer boost in a very convenient way, while remaining at about -5 or lower with our AVR trim.
It should be emphasized that there is no particular reason not to just use the gain control on a sub to add volume post-calibration. For people wanting to add really substantial bass boosts--up to, or in excess of 10dB or 12dB, some gain increase, in excess of the original gain setting, is generally necessary anyway, in order to achieve the bass boost desired by the user.
* [It should be noted that some Denon AVR's have a feature called "Subwoofer Level Adjust". When this feature is used, the subwoofer trim level is reset to 0.0, and the adjustment is made on top of that. So, starting at -11.5, post-calibration, and adding 5dB of boost with that feature, will actually result in a net trim level of +5.0 in trim, instead of -6.5 in trim. That is a net increase of +16.5dB instead of +5dB. It is highly recommended to turn that feature off, and instead to make any necessary subwoofer volume adjustments either with the Channel Level Adjust in the Audio menu, or with the trim controls in the Speaker: Manual: Test Tone area of the Denon AVR.
Edit: Apparently, this glitch has been fixed in newer Denon models, and with recent Denon firmware updates. If your AVR does not have an on/off switch for the Subwoofer Level Adjust, using that feature will add the correct amount of subwoofer volume, just as would be the case if you used the test tones. If your AVR does have the on/off switch, you can test the SLA feature to determine whether it is adding the appropriate amount of subwoofer boost.]
[It should also be noted that ARC Genesis doesn't allow users to ignore the gain setting zone in the way that Audyssey does. So, using high gain settings to achieve low-trim settings, doesn't work well with the most recent version of ARC. When using ARC Genesis, users should simply follow ARC's instructions during the Quick Measure process. After ARC has run its automated calibration routine, users can still turn-up the gains (symmetrically) on their subs, if necessary, in order to achieve more bass, or to reduce high AVP trim settings. It is always going to be a matter of listener discretion as to whether or not to avoid high AVR/AVP trim levels. As with any subwoofer/AVR combination, there may or may not be audible clipping with high AVR trim levels.]
To continue the general discussion of where to add subwoofer boosts, the usual recommendation to employ the AVR trim is more a matter of convenience and of accuracy than one of necessity. Some subs don't have digital gain controls, for instance, so fine-tuning the gain can be more difficult, as can on-the-fly adjustments during a particular movie, or music listening session. And, it gets even less convenient when multiple subs are connected together, or when gain controls are difficult to access easily. Using the trim controls in an AVR allows for very convenient and precise adjustments in sub volume. But, the most important thing is to make sure that the real boost comes from the subwoofer amp, and not just from the AVR, whichever adjustment method is ultimately employed.
Assuming that some initial sub boosts are to be accomplished using AVR trim, then starting with a low trim level post-calibration would be helpful. A low trim level might be defined as -9 to -11, but not exceeding -11.5 in Denon/Marantz units. (With other manufacturers, just determine what the minimum trim level settings are in order to ascertain what your optimum low trim setting should be.) As stated earlier, it may take an SPL of 78dB, or higher, to achieve that optimum low trim level. However, it is important not to go lower than -11.5 in trim, in Denon/Marantz units. (For example, I believe that the lowest trim levels we should calibrate our audio systems to are -14.5 with Onkyo, and -9.5 with Yamaha, based on their respective trim level limits of -15 and -10.)
If a trim level of -12 is set, with Denon/Marantz units, there is no knowing what the actual volume of the subwoofer is. The AVR simply ran out of negative trim at -12. The actual sub volume might be 80dB, or even 85dB. If so, you might not like the way it sounds to have your sub so much louder than the rest of your system. And, you would not have an easy way to turn down the subwoofer volume, if your trim level were already at the lowest setting. You also could never be sure what your actual sub volume is, and as a result, you could find yourself running out of headroom sooner than expected. So, for instance, you want a negative trim setting not exceeding -11.5 in Denon/Marantz units.
Think of the process of adding a sub boost this way. When you raise the gain level in the sub, so that the sub produces more than 75dB at the MLP, you are making a deposit in the bank, of amplifier power from the sub. So, for instance, let's say you start with a trim level in the AVR of about -9 to -11. Now, you can withdraw amp power from the bank, using your AVR trim control. You would, for instance, do that by increasing your trim setting to about -6 or -5. As noted earlier, a +3 to +6dB boost would be pretty typical, even with DEQ engaged. But, there is no free lunch. As you begin to approach 0.0, the bank deposit of amp power that you made with the higher gain setting is used up, and now you are using AVR amp power, which as noted, is not as powerful. Using AVR amp power can, in some instances, result in clipping (distorting) your subwoofer(s) or it can, in some cases, result in undesirable mechanical noises.
[Listeners sometimes mention the importance of using only the AVR trim to add subwoofer boosts, because they are able to know exactly how much additional SPL they have added that way. I have never been convinced of the importance of that. As a general rule, we are simply adjusting our subwoofer volume to match our personal listening preferences, anyway.
If we suspect that we may be running out of headroom, that is a separate question, irrespective of the precise amount of subwoofer boost we are adding. If we have ample headroom, then I'm not sure that the precise amount of SPL we have added is really very important. Some subwoofers add about 1.5dB per click, where there are detents in the analogue dial. Others may add about +3dB. The subwoofer maker can usually confirm the amount per click if listeners are really interested in knowing approximately how much SPL each click represents. And, of course, listeners can always measure their subwoofer SPL, as they add gain, if they want to know more precisely the amount of subwoofer boost added post-calibration.
Some listeners also wonder how important it is to keep track of how many clicks they have added post-calibration, and are concerned about not being able to do that with analogue dials which don't have sufficiently fine detent markings on the outside of the dial. I know that some listeners have used washable magic markers, or small pieces of tape, to mark the original post-calibration setting, before they add any additional gain boost. We can be pretty creative about that kind of thing if we are really curious.]
Section II-D: Master Volume Levels And Sub Boosts
There is a relationship between subwoofer volume and master volume (MV). As your MV increases, the subwoofer volume goes up correspondingly, and more demands are placed on the sub. It is important to remember that the subwoofer is not only playing the LFE channel, but also providing bass support for all of the other channels in a typical HT system. So, as the MV increases, the demands on the sub go up much faster than for the other channels, particularly in a movie with a lot of low-frequency content. It is worth noting that 5.1 movies (and some bass-enhanced music) can have very low frequency content in all of the channels, and not just in the LFE (low frequency effects) channel. The subwoofer has to (and should) play all of that low frequency content.
It is recommended by a number of subwoofer experts, two of whom are quoted in the FAQ, that it is advisable to keep sub trims well in the negative range (below 0.0). That is particularly important as MV's approach, or exceed, -10. In Denon/Marantz units, that is 10dB below Reference (or 70 on the absolute scale) in your AVR master volume. Both of those experts quoted in the FAQ, Ed Mullen of SVS, and Mark Seaton have, subsequent to the entries in the FAQ, recommended staying well in negative trim levels, period. To follow their advice, and to avoid the possibility of distortion, we would want to keep our trim levels in about the -5 range, or lower, at even moderate listening levels. Again, that is easy to do by simply raising the gain on the subwoofer(s).
* [In addition to the possibility of clipping the subwoofer signal, with higher AVR trim levels, there is another potential reason for keeping gain levels relatively high. It is addressed in the following section titled Gain Settings and Maximum Sub Output.]
High sub gain levels, which still result in high trim levels, are indicative of a sub which is under-powered for the space, and/or the distance from the MLP. It could also be indicative of a specific placement problem, where either the sub or the MLP is located in a null. In the first instance, the only remedy would be a more powerful sub, or multiple subs, or a different (probably closer) sub placement. In the situation where the sub or the MLP were located in a null, a subwoofer crawl should be done to determine proper sub placement. Although subwoofer placement is not a direct part of this particular discussion, it is a very important factor in sub performance.
If you never intend to approach about -15 MV or higher, then the advice to set your sub gain high enough to obtain a strongly negative trim level might be less important. (Even then, however, that could be somewhat dependent on the use of DEQ and/or the use of independent sub boosts.) And, if you don't believe that you will ever want to listen at high volumes, or to boost your subs, then starting with a trim level of about -5 or -6, should be perfectly fine.
But, most people on this and other threads seem to average at least a +3 to +6dB bass boost after calibration, and some people add much more than that. When DEQ (with its own bass boost) is not employed, boosts of +12dB, or even more, are not uncommon. So, the advice you will most commonly see on this thread is to start with a negative trim setting of about -9 to -11 post-calibration, in order to maximize your ability to add sub boost, with your AVR trim control, while still using the sub gain you deposited in the bank.
Although this advice is not consistent with the explanations and recommendations in the FAQ, the more current advice supersedes the older advice in the FAQ. I would personally recommend following the advice to maintain a negative sub trim, preferably of -5 or lower, as a matter of best practice, even if you believe that you will never approach -10 or -15 MV. (As noted in the section just below, keeping a relatively higher gain level may work to your advantage, in any event.)
There is no telling who might, inadvertently or otherwise, run the volume control up on your system, or when unexpected peaks in very low bass (in electronically-enhanced music, or in movies) might cause some distortion to occur. And, if your sub happens to be approaching its max output limits, even at lower master volumes, the lower trim level would provide an additional measure of confidence that you weren't clipping the subwoofer signal.
While it is unlikely that most good modern subs would be damaged by a bit of distortion, or by an inappropriate use of AVR amp power, I know of two well-documented instances of a JTR Orbit Shifter, which is an extremely powerful and well-made sub, frying a voice coil (due to overheating) just from playing electronic music, downloaded from YouTube, at a very high volume, with a high AVR trim level. (Both instances involved the same user, who didn't quite remember the lesson from the first instance.)
And, even if no damage could ever be done as a result of clipping the sub signal, listening to distorted bass is sort of antithetical to the whole idea of good sound quality, and of using automated room EQ to achieve it. Clipping also consumes another +3dB of amplifier power. Where each +3dB of SPL equals a doubling in amplifier power, that +3dB increase, due to clipping, is significant. When proper gain/trim protocols are followed, it is also less likely that inappropriate noises, such as port chuffing, or of drivers hitting limiters, could occur prematurely. So, an ounce of prevention is worth more than a pound of cure, in this case.
Again, you can use a combination of increased subwoofer gain, and some increase in AVR trim, to raise the volume level on your sub to any level you choose, while still maintaining an AVR trim of about -5, or less. (Since originally writing this guide, I have seen even more recommendations from subwoofer makers to be at -5 or less in the AVR trim.) That will help you to avoid the possibility of clipping the subwoofer signal. And, raising the gain control on the sub(s) post calibration, will have no effect at all on the way that Audyssey EQed your system.
Section II-E: Gain Settings And Maximum Sub Output
There is another aspect to the gain/trim issue that is worth mentioning. Depending on how the DSP in a given subwoofer is implemented, the subwoofer may only be able to achieve max output levels with the gain control set very high. Some subwoofers are only able to achieve maximum output levels when the gain control is set to, or very near, the highest setting. So setting a lower gain control, and a correspondingly higher AVR trim control, might not result in the same amount of peak bass SPL, irrespective of issues of clipping.
Apparently, this issue may be more common in subwoofers with digital (rather than analogue) controls. But, according to several examples I have observed from various threads, the issue is not at all limited to subwoofers with digital controls. Some subwoofers with analogue controls may have the same issue of not being able to achieve higher max output levels with low gain settings.
How important this max output issue actually is probably depends on the situation. For instance, I believe that a relatively lower gain setting might cause a ported subwoofer to chuff prematurely. Again, depending on the situation, even someone who is listening at a fairly moderate listening level, let's say -15 or -20 MV, might experience issues if he were using a significant subwoofer boost, either independently or on top of DEQ.
Putting a sudden peak demand on the subwoofer, with the right low-frequency content, might not enable the subwoofer to access the full output that it is designed to produce, if the gain level isn't fairly high. In that case, the subwoofer just wouldn't play the low-frequency content at the volume it was supposed to. In other words, it would simply stop getting any louder during that peak content. Whether we would even notice that, or whether we would hear the subwoofer make any audible sounds of distress, are separate questions.
But, unless we are sure that our subwoofers can achieve max volume levels with low gain settings, it is probably a good precaution to keep gain levels fairly high. Typically, that will mean using corresponding lower AVR trim levels for our subwoofers. This is not an issue that I have often heard addressed by subwoofer makers, but I suspect that many would intuitively know that some subs produce max volumes only with high gain levels.
This is just speculation on my part, but I think one reason that this issue isn't discussed more is because sub makers are not typically testing their subs as part of a calibrated HT system. So, they aren't dealing with gain/trim relationships at all in their design and testing process. When they want to push one of their subwoofers to its limits, or they want to measure its maximum output, they just max out the gain control on the subwoofer itself. It is only when subwoofers are calibrated as part of an HT or audio system, with an inverse relationship between gain and trim, that this becomes an issue. But, I believe that it can be an issue, and I believe that is another potential reason for attempting to keep gain settings fairly high.
CEA 2010 testing, performed by Data-Bass, always measures max output with gain controls at the maximum setting. And, as noted, some subwoofers may not be able to produce those same max SPL numbers, that we see on Data-Bass, or from other professional sources, with lower gain settings. This won't be true for all subwoofers, but as a matter of best practice, I believe that it may be generally advisable to keep gain settings fairly high, and AVR trim settings fairly low, in order to maximize available headroom. An exception to this general policy could be a situation where a lower trim setting didn't successfully power a subwoofer on, when it was set to Auto On mode. But, that would be extremely unusual with most receivers and processors.
* People with some Yamaha AVR's are apparently much more likely to experience issues with subwoofers not turning on automatically unless AVR sub trim levels are relatively high--perhaps even fairly close to 0.0. That is due to the lower voltage signal sent from some Yamaha AVR's to the subwoofer. In some cases, this may have been due to some defective sub outs on some Yamaha AVR's, which were replaceable under the Yamaha warranty. Yamaha AVR's from about 2017 on were previously reported to have addressed the problem, but that doesn't seem to always be the case. Both Yamaha and Onkyo AVR's may also experience a calibration (level-matching) problem due to low voltage signals to the subwoofers. That issue is addressed in the last few paragraphs of this section.
If subwoofers will not turn on automatically in Auto mode, without higher AVR trim levels, then the higher trim levels may be slightly less likely to lead to clipping issues, since the voltage from the AVR was lower to start with. Some Yamaha owners use a Y-connector into both subwoofer inputs in order to double the voltage going to the sub. And, that typically resolves the Auto On issue. Of course, Yamaha owners can also choose to just leave their subs on all the time, if the Auto On issue proves to be a real problem. That will consume slightly more energy, but will not affect the operation or longevity of the subwoofer.
AVS member @Basshead recently mentioned a clever solution for achieving more headroom, with higher gains and lower trim levels, which seems to circumvent the Auto On problem with Yamaha AVR's. He went from a -1.5 subwoofer trim level to a -4.5 trim level, with a comparable gain boost, and obtained +3dB more headroom, prior to clipping. But, his subwoofer didn't power-on reliably when watching TV at very low master volume levels. So, he lowered the trim levels on all of his other channels by -3dB, and raised his MV level by +3dB, and is now able to have his sub power-on reliably for low-volume TV content, while still having more headroom available for louder movie viewing. This is an additional technique that Yamaha owners might wish to try.
One final issue involves Yamaha AVR's which yield abnormally high trim levels no matter how high the subwoofer gain levels are turned-up. This can be a much more significant problem than the auto-on issue. In a recent example from early October of 2021, a new Yamaha RX-A8A exhibited this issue of setting abnormally high trim levels with the subwoofer gains also set high. Some before-and-after screen shots of trim levels, showing the problem being solved with Y-connectors, are illustrated on Page 218 of the Guide thread. If you believe that your Yamaha (or Onkyo) AVR may be exhibiting similar behavior, it could be worthwhile to look at the pictures on Page 218.
Apparently, this can also be an issue with Onkyo AVR's. The typical voltage sent from AVR's to subwoofers is about 2.0V or slightly higher. I'm not sure what the voltage coming from some Yamaha AVR's is, but I have been informed by AVS member @fattire that with some Onkyo's it is only 0.9V. If the subwoofer receives substantially less voltage than the typical 2.0V, then its performance may be adversely affected, and that limitation may show-up during calibration. The subwoofer trim levels may be set too high even with very high subwoofer gain levels. That can be an indication that the subwoofer is not able to reach its normal output potential.
Where this problem is believed to be occurring, the way to correct it is to use a Y-connector into both sub inputs. That will double the voltage going from the AVR to the subwoofer. So, using Y-connectors can allow subwoofers to turn on automatically in some cases, and they can also be used to allow the subwoofer to achieve its full operating performance. Once again though, the Y-connectors are only effective where the voltage coming from the AVR is insufficient. They won't improve on the inherent performance capabilities of the subwoofer. Its own amplifier will determine the subwoofer's inherent capability.
This is an example of the type of Y-connector which would be used to increase the voltage going from the AVR to the subwoofer:
Amazon.com: Amazon Basics 2-Male to 1-Female RCA Y-Adapter Splitter Cable - 12-Inches
Amazon.com: Amazon Basics 2-Male to 1-Female RCA Y-Adapter Splitter Cable - 12-Inches
www.amazon.com
Section III: Setting Crossovers:
This general discussion of bass-management, and of setting crossovers, applies to other systems of automated calibration and not just to Audyssey. The same questions come up so many times that I think it is worth emphasizing some of the basic crossover concepts. We use crossovers between our speakers and our subwoofer(s) in order to bass-manage our audio systems. In audio systems where there is no subwoofer, there will be no bass-management required, and speakers will always be set to Large or Full-Range.
The Subsections in Section III are as follows:
III-A: Crossovers from Speakers to Subwoofers
III-B: Low Frequency Effects Channel
III-C: Cascading Crossovers
III-D: Bass Localization
III-E LFE+Main
Section III-A: Crossovers From Speakers to Subwoofers:
Subwoofers can be used with two speakers in a stereo system, or they can be used with a 5.1, or larger, audio system. Whenever they are used, it is necessary to have a way to cross over from the speakers to the subwoofers, so that the subwoofers can play bass content below a designated frequency. Good subwoofers are designed for the sole purpose of playing bass frequencies below about 150Hz.
(Subwoofers will typically have a setting labelled "LPF" (low-pass filter) or "Crossover". It may be an analogue knob on the amplifier plate. That setting should generally be at the highest setting, which will usually be 150Hz. That will allow the subwoofer to play frequencies up to 150Hz, before it starts to roll-off. An exception to that general rule is explained in Section III-C: Cascading Crossovers.)
As frequencies go below about 120Hz, and especially below 80Hz, subwoofers typically perform their specialized function much better than even large tower speakers can. We set crossovers to allow our subwoofers to take over duties below a selected frequency. That selected frequency will depend to some extent on the native capability of our speakers, and it will depend somewhat on the speakers' placement in the room, since room placement will strongly affect low-frequency performance of any transducers in both positive and negative ways, as explained in Section I.
Section I also noted the importance of placing tower or bookshelf speakers where they can point toward the listener, so that mid and high-frequencies will be heard clearly. The advantages of spreading speakers apart to achieve a wider soundstage, and potential issues with early reflections from side-walls were also discussed. But, those are all issues affecting mid and high-frequencies. And, they may necessarily dictate the placement of our front speakers, if we want to maximize our sound quality. The room geometry and the furniture arrangement in the room, may also be factors in the positioning of our front speakers.
Bass frequencies have different issues with respect to placement though. So even if our front speakers are not optimally positioned with respect to bass frequencies, our subwoofer(s) may be able to compensate for that. Good subwoofer placement can be completely independent of the placement requirements for our tower and bookshelf speakers. This is why, even those of us with very capable speakers, may wish to have subwoofers which we try to strategically locate for optimal bass performance, and which we bass-manage for optimal integration with our speakers. That is especially helpful for movies, where the low-bass demands can be very significant.
When we look at a 5.1, or larger, system, we see even more importance attached to the subwoofers. First, the subwoofers must provide bass support for all of the regular channels, just as they would in a two-channel system. In 5.1 content, those regular channels can have bass peaks up to 105dB, depending on master volume and bass levels, and the frequencies can go very low at times. Second, the subwoofers must play all of the content in the LFE (low frequency effects) channel, at peaks of up to 115dB, also with potentially very low-frequencies. Subwoofers which are powerful enough for the room, and for the individual listener's preferences, perform this double duty just as they are designed to.
Those two different subwoofer functions are controlled by two different mechanisms. As explained below, the bass content redirected from the speakers is controlled by crossovers. The LFE content is controlled by a low pass filter, called the LPF of LFE. That LPF setting is explained in more detail in Section III-B. The default LPF setting in most AVR's is 120Hz.
The low-bass content in the regular channels is controlled by crossovers set for each pair of speakers (or for the center channel). In order for bass to be redirected from the regular channels to the subwoofer(s) speakers must be set to Small, with a crossover. The assignment of crossovers for each channel is accomplished during the initial calibration process, and then may be modified by the user. Crossovers may be assigned globally (such as 80Hz for all channels) or they may be assignable individually, for speakers pairs and for the center channel, depending on the specific AVR.
As stated earlier, where we have subwoofers in our audio systems, and wish to employ them as they were designed to be used, some system of bass-management is necessary to split the signal between the speakers and the subwoofer, so that the subwoofer can handle low-frequency content in the regular channels, while the speakers continue to play all of the other content. And, that split can only be accomplished through a setting of Small, with a crossover. Determining where that split between the speaker and the subwoofer should occur starts with the calibration process, where initial crossovers are assigned by the AVR. And, it continues after the calibration process, as listeners adjust their crossovers to achieve their specific listening objectives.
The frequency at which a signal split should occur may be different for different speakers in our audio system, depending on their low-bass capabilities and on their placement in the room. The subsection on room gain (in Section VII-B) explains something of the unpredictable ways that a room, and placement within a room, can affect the bass response of a subwoofer. The same explanation applies to speakers. Room placement can affect a speaker's bass response, which in turn determines where a crossover will be set. (Where crossovers must be set globally for all of the speakers in an audio system, as with Yamaha AVR's, some compromise may be necessary.)
It was noted in an earlier section that distances (timing) and channel trim levels are determined by microphone position number 1. That is not the case with crossovers. Crossovers are set based on the fuzzy-weighted average of the frequency response from all six or eight microphone positions. As a general rule, crossovers in a full calibration will not vary much from where they would be set based on the first few mic positions. But, they may vary slightly, once Audyssey has calculated the FR at all available mic positions.
Audyssey (and other systems of automated calibration) accomplish bass-management during the initial calibration. When Audyssey measures all of the speakers in an audio system, it reports the measured F3 point of each channel to the AVR or AVP. (The F3 point is the frequency where a speaker is reaching the bottom of its frequency response, and rolling-off in SPL by 3dB.) That point at which a speaker begins to roll-off naturally by -3dB will be dictated by both the inherent capability of the speaker, and by its position within the room.
Once Audyssey has completed its measurements of frequency response for each speaker, the AVR then sets that speaker, or speaker pair, to either Small or Large (also called Full-Range), based on its own internal programming. If a speaker begins to roll-off in the mid to high 30's (or lower) the speaker will be set to Large. At any frequency above about the high 30's, the speaker will be set to Small, and a crossover will be assigned. The weaker of two speakers in a pair will control the crossover, as Audyssey is specifically designed not to EQ below the F3 point of any speaker, or subwoofer. Speaker location, with respect to boundary walls and room modes, may make one speaker in a pair roll-off earlier than the other one.
As stated, if a speaker's F3 point is somewhere in about the upper 30's, the AVR will round-up, and set that speaker's crossover to Small with a 40Hz crossover. Crossovers will always round upward, so an F3 point of about 42 or 44Hz would round-up to a 60Hz crossover, and an F3 point a little above 60Hz would round-up to an 80Hz crossover. The exact number used to round upward would probably vary somewhat among different AVR makers, but the basic principle involved is applied in both Audyssey and non-Audyssey systems. From the user's standpoint it is important to note that, without independent measurement, there is no way of knowing exactly where a speaker's roll-off actually occurred. So, it may be advisable to be conservative with crossover settings, after a calibration.
[Sometimes people observe changes in crossover settings that seem to coincide with a change in the type or number of subwoofers. Crossovers may vary slightly from one calibration to the next, and certainly can change due to relatively small shifts in speaker positioning. As explained in the later section on room gain, boundary gain due to proximity to walls can affect the low-frequency response that Audyssey is measuring, as can specific room modes. Small changes in microphone positioning between calibrations can also affect crossovers, as Audyssey is averaging the results of all of the mic positions performed in a calibration. But, Audyssey is only measuring each speaker in isolation, without reference to subwoofers. Any change that appears to coincide with some change to a subwoofer is entirely coincidental.]
The initial setting of Large, or of Small with a 40Hz, or 80Hz, or higher crossover, does not constitute a recommendation, either by Audyssey or by the AVR. This is an important point to understand. The initial crossover setting is simply a modest setting designed to somewhat protect the speaker, while providing information about that speaker (or speaker pair) to the user. The information the initial crossover setting provides tells the user something about where a particular speaker is actually rolling-off by -3dB, at that particular position in the room, and informs the user that no EQ is being performed below that approximate point defined by the crossover. It is then the user's responsibility to interpret that information, and to decide whether to leave the crossover at that initial setting, or to change it.
* It has already been noted that setting speakers to Small with a crossover is the appropriate way to direct lower bass frequencies from the regular channels to the subwoofer(s). As a general rule, I would suggest that where a calibration sets crossovers to 40Hz, an increase to at least 60Hz, or perhaps even to 80Hz, might be advantageous. I believe that the 40Hz crossover setting covers a very narrow range of frequencies from about 36-38Hz to about 42-44Hz. Anything higher than about 42Hz or 44Hz will probably automatically round-up to a 60Hz crossover. With speakers already rolling-off at about 40Hz or so, a good subwoofer should handle that 40Hz frequency, and those at least a half-octave higher, much more effectively. I believe that 80Hz would typically be a much more conservative setting for most speakers.
This might be a good opportunity to give a practical explanation for why it can be advisable to raise crossovers after the calibration process is complete. Let's take the example cited above, of an initial crossover setting of 40Hz. If we see a crossover setting of 40Hz, that may confirm our belief that a speaker (or speaker pair) is pretty capable. But, assuming that a speaker is already down in volume by -3dB, at about 40Hz, what does that really mean in practical terms? It means that at 75dB, the speaker is already running out of gas at about 40Hz. And, 75dB isn't very loud, for bass frequencies, in 5.1 content.
If someone is listening at a master volume (MV) of -20, in a calibrated HT system, that speaker will need to be able to play peak volumes of 85dB with 5.1 content. At -15 MV, that speaker will need to be able to play 90dB for peak volumes, with 5.1 content, and so on as the volume level increases. We already know from previous sections that we can't hear lower bass frequencies as well as those in our normal hearing range, and at -15 MV, with a 40Hz crossover, 40Hz is going to be playing about-18dB softer than it should be at a master volume of -15.
Asking it to play frequencies that it really can't play effectively shouldn't hurt it. That's why the high-pass filter in the crossover makes it play softer once it gets to that F3 point. But, at best we simply won't hear the 40Hz frequencies, and at worst we may hear some distortion, compression, or clipping from that speaker. Alternatively, if we had crossed that speaker at 80Hz, instead of at 40Hz, the subwoofer(s) would have been able to play the 40Hz, and 50Hz, and 60Hz frequencies, much more easily and with much less potential distortion or compression. At a minimum, the sound should be more balanced and the lower frequency sound quality might be clearer too.
Here is another factor we should also consider, if we have something such as Audyssey's DEQ, which is pre-programmed system of loudness compensation. YPAO (on Yamaha AVR's) has something similar to DEQ. DEQ is explained in detail in Section V-A, but briefly, DEQ boosts the bass in all of the channels as listening levels drop below the Reference volume of 0.0. DEQ adds approximately +1.1dB, for frequencies between 70Hz and 120Hz, for each -5dB decrease in master volume. And, it adds progressively more bass boost below 70Hz to a total of +2.2dB at 30Hz and below. At 40Hz, DEQ would be adding about +2dB to each channel for every -5 MV.
So, using our previous example of -15 MV, DEQ would add another +6dB at 40Hz. And, at -20 MV, DEQ would add another +8dB to a speaker which was already trying to play 13dB louder than it actually could. Understanding how DEQ can put additional demands on the low-frequency performance of our speakers may help to explain why some people feel that DEQ makes their audio systems sound bloated. When we talk about compression of bass frequencies, we mean that the lowest frequencies stop getting any louder, while the mid-bass frequencies start to dominate more. That could result in what is often described as boomy or "one-note bass".
None of this means that we shouldn't use DEQ, or that we can't enjoy our speakers to the fullest extent. But, it does mean that we need to understand what an initial crossover setting actually tells us, so that we can help our speakers to play with as little strain as possible. And, that's why most of us added subwoofers to begin with. We wanted to be able to play even lower frequencies, with more volume, than we could obtain from our other speakers. Letting the subwoofers do what they are designed to do may result in better overall sound quality.
Of course, bass localization can be an issue with higher crossovers, depending on where a subwoofer is located, so we may have to balance our interests. A technique that makes subwoofers roll-off faster, above crossovers, is called cascading crossovers. That can help with bass localization, and it may also help with overall bass clarity. Cascading crossovers are explained in Section III-C. And, bass localization is described in some detail in Section III-D.
[FWIW, as noted previously, I think that an initial crossover setting of 60Hz may be especially problematical. Assuming that our AVR is rounding-up from anywhere in the low-40's to about 60Hz or so, then it can be very difficult to know where the F3 point of our speaker actually is. A speaker might be starting to roll-off at about 42 or 44Hz, or it might already be down -3dB in SPL somewhere in the mid to upper-50's. In the absence of measurements to tell me where my actual F3 point is, with a 60Hz crossover setting, or a strong preference for the sound with it set that low, I would personally be more comfortable raising the crossover to at least 80Hz.]
With crossovers already set to 80Hz during the calibration, users might also wish to experiment with slightly higher crossovers, or they might just leave them set at 80Hz. The lower bass frequencies put more demand on a speaker, or a subwoofer, than the mid-bass frequencies do. And, that demand can create distortion in our speakers, particularly at higher volume levels.
I have tried to think of a good rule-of-thumb to use when trying to decide whether or not to be more conservative with our crossovers. I have already recommended increasing crossovers for 40Hz and 60Hz initial settings. It seems to me that master volume levels of about -15, for 5.1 movies and bass-heavy music, would be a pretty safe dividing line for an 80Hz crossover. People listening at about -15, would probably be okay leaving an 80Hz crossover, which was set by the AVR, at that 80Hz frequency. (Using DEQ might also work better with crossovers of at least 80Hz, as DEQ only boosts the frequencies between 70Hz and 120Hz by approximately +1.1dB per -5 MV.)
People listening above about -12 MV, though, might want to raise the crossover slightly higher than that initial setting of 80Hz to relieve any potential strain on the speakers. Remember that bass frequencies will consume more of a speaker's total headroom than higher frequencies will, so the louder we play, the more conservative we might want to be with our crossovers. An exception might involve speakers with very high sensitivity ratings. (Spoiler alert: some manufacturer's may inflate those sensitivity ratings, partly by not showing at what distance the measurement is taken.)
I think it typically makes sense to be a little more conservative, with almost any of our speakers, below about 60Hz or 80Hz. Different speakers, different rooms, and different speaker/subwoofer interactions, however, could all influence the selection of a crossover. Actual experimentation, with careful listening and/or measurement, may be required to make final setting decisions, and the selections may depend heavily on individual user preference. I think that most of us will hear it if our speakers are consistently having trouble playing the material we like, at the volumes we enjoy. In that case, higher crossovers, or lower listening levels, may help to reduce that slightly audible distortion.
Again, it's one thing for a speaker to be able to hit 75dB, at let's say 40Hz, without audible distortion or clipping. It may be a very different thing if that same speaker attempts to hit that 40Hz frequency at 85dB or 100dB. This is why it may make sense to be a little bit conservative with our crossovers.
**
[Listeners who are curious about the specific capabilities of their speakers may want to investigate on their own, via measurements. But, in the context of setting crossovers, this may actually become a little more complex than it seems. Although a listener might choose to try to measure his speakers with a test disk and an uncalibrated SPL meter (such as a Radio Shack meter), I don't think that the results would be very reliable, compared to what the calibrated microphone of an AVR can already do. Uncalibrated SPL meters are notoriously unreliable for lower bass frequencies.
I think that really accurate results would probably require the use of a calibrated UMIK-1 and REW, or some comparable measurement system. The tester would want to see where his speaker was rolling-off by 3db, via a frequency response graph, with the volume at sufficiently loud levels. Or, he could do a compression test, to determine the same thing. If he were using DEQ, he should have it on for these tests. And, he could listen to the audible results during those tests, to identify distortion, clipping, or compression. He could then correlate the results to his typical listening levels.
The following link to the REW thread will help users, so inclined, to understand what is involved in the use of REW:
Simplified REW Setup and Use (USB Mic & HDMI...
For most HT owners, relying on the AVR to have correctly identified the speakers' roll-off points, via the initial crossover setting which takes that roll-off into account, is probably going to be sufficient. Then, understanding that the AVR has identified the roll-off points for us, and has set crossovers accordingly, we can exercise independent judgment on whether to leave the crossover setting where it is, or to raise it. The entire purpose of this subsection is to assist in understanding what the AVR is actually doing, when it sets crossovers, and to assist in the subsequent application of independent judgment in deciding what to do with those initial settings.]
As an aside, the reason that the Large or Full-Range setting is still necessary in modern audio systems is because not every system has a subwoofer, and not every user employs subwoofers for all listening material. For instance, some listeners may choose to listen to a music genre (which may have relatively little low-bass content) with their speakers set to Large, and without any subwoofers engaged. But, in order to employ a subwoofer for anything except the LFE channel, it is first necessary to set speakers to Small with a crossover.
[Some AVR's also have a feature which allows a Large setting with subs employed. The setting is called LFE+Main, or Double-Bass, and is explained in some detail at the bottom of this section, in Section VII-E. That setting may increase the apparent quantity of bass, but may also introduce considerable distortion in the process. It is not generally recommended by Audyssey and others, from an audio quality standpoint, although that is strictly a user-preference issue.]
It should be noted that owners are often surprised by the crossovers set by their AVR's. Sometimes, they are surprised that the crossovers are set so high, and sometimes, they are surprised that the crossovers are set so low, because in either case the crossovers don't align with their expectations. Two major factors contribute to that surprise. First, speaker makers frequently inflate the low-frequency specifications of their speakers. Second, room placement plays just as important a role in the low-frequency performance of our speakers as it does for our subs. The optimum location for a particular speaker (or speaker pair) may give it extra low-frequency response, due to boundary gain or due to favorable room modes at particular frequencies, or it may rob it of some low-frequency performance. Audyssey and other systems of automated calibration and room EQ will simply measure what they detect, and report that to the AVR, which will set the crossovers in accordance with its own algorithm.
As a general rule, crossovers can always be set higher than where they are set automatically by the Audyssey calibration. This is a user-preference issue, and may depend on what sounds (or measures) best to a particular individual. It is not a good idea to set crossovers lower than where they were set automatically during calibration, because those speakers will not be receiving any benefit from room EQ, somewhere a little below the original crossover point. They will also be down -3db in measured SPL at the point where the EQ stops. And, they will continue to play softer and softer as the frequencies go lower, so, they will not be providing much audible benefit at that point, anyway.
In addition, running speakers with crossovers below the original calibration setting, may consume valuable amplifier power, and may result in some distortion, while the more powerful subwoofers are being correspondingly under-utilized. So again, it is typically better not to reduce crossovers from wherever our AVR's set them. Crossovers of at least 80Hz are typically recommended in THX standards, and for general best-practice purposes, as subs will nearly always do a better job of reproducing the mid-bass frequencies up to 80Hz or so. 80Hz is also used as a standard frequency for where many people will not be able to localize a subwoofer.
*** Sometimes, HT owners may be a little reluctant to set crossovers of 80Hz, or higher, due to concern that they won't be using the full capabilities of their tower, or large bookshelf, speakers. But, in considering that, it is important to understand how the crossovers in our AVR's actually work. Crossovers are not like brick walls, where the speaker suddenly stops playing everything below 80Hz, and the subwoofer suddenly starts. We might be able to hear that kind of abrupt transition from speaker(s) to subwoofer(s).
Instead, when we set a crossover, the AVR implements a high pass filter (HPF) for the speaker(s) and a low pass filter (LPF) for the subwoofer(s). The high pass filter is designed to pass all of the frequencies above that point, while the low pass filter passes all of the frequencies below that point. But, both filters have a slope which gradually reduces the volume of the speaker or subwoofer to insure a smooth transition at the crossover point. (The filters also gradually reduce the volume to insure that neither speakers nor subwoofers are damaged by trying to play frequencies that they shouldn't be playing, at anything approaching full volume.)
The HPF for the speaker(s) is typically a 2nd order filter with a slope of -12dB per octave. The LPF for the sub(s) is typically a 4th order filter with a slope of -24dB per octave. In theory, at an 80Hz crossover, the speaker would be playing at -3dB, and the subwoofer would be playing at +3dB. That would maintain equilibrium at that frequency, so that there would be no apparent change in volume there, while allowing both transducers to start their roll-offs on the other side of the crossover.
(As is noted elsewhere in the Guide, at somewhere near the crossover is also where there can sometimes be phase-cancellation between speakers and subwoofers since they are playing the same content at about the same volume level. where that occurs however, the area of cancellation is often quite small, and there is no audible effect. It should also be noted that strong subwoofer boosts can raise the area where phase-cancellation might occur, since the subwoofer is playing much louder than the speaker at 80Hz, for instance, but closer to the same volume at 90 or 100Hz. This is not an issue that we should normally be concerned about unless we hear something that we shouldn't. Once again, our personal hearing preferences should be our final judge of which crossovers to use.)
Putting the previous example of how crossovers work into practical terms, with an 80Hz crossover, a speaker would still be playing 60Hz content, but a little softer than it plays 80Hz content. It would play 60Hz at about -8dB, and 40Hz content about -12dB softer than it was playing at just a little above 80Hz. The frequencies between 40Hz and 80Hz would constitute one octave (40Hz times 2).
[Note: An octave, in this context is a series of 8 notes, where the frequency of one is twice that of the other. So, from 40Hz to 80Hz would be an octave. From 80Hz to 160Hz would be one octave.]
Looking at that same 80Hz crossover, from the standpoint of the subwoofer, the sub would still play 100Hz and higher frequencies, but at gradually decreasing volumes. By the time the it reached 160Hz, it would play -24dB softer. The subwoofer's -24dB octave would consist of the frequencies between 80Hz and 160Hz (80Hz times 2). Once again, strong subwoofer boosts would slightly affect the attenuation of bass above the crossover. In general, the crossover filters, set by the AVR, allow the speaker to gradually give way to the subwoofer, while allowing both subwoofer and speaker to play a little above/below the selected crossover point. It's all about trying to achieve a more blended transition.
The crossovers within our speakers are similarly designed to create gradual transitions from woofer to mid-range, and from mid-range to tweeter (in a three-way speaker), although the precise amount of attenuation employed in the filters may vary among different speaker designs. When employed properly, crossovers enable us to listen to our audio content with no audible transitions at all from driver-to-driver within a speaker, or from speaker-to-subwoofer within an audio system.
As far as setting crossovers within our audio systems is concerned, there may be circumstances in which a crossover even lower than 80Hz is desirable, where measurements, or our own perceptions of sound quality, guide us. A situation where our speakers were set to Large during the calibration (because the measured F3 point was in the mid-thirties or lower) might lend itself to setting a crossover lower than 80Hz. For instance, in some cases, it is possible that a 60Hz crossover might provide an apparently smoother transition, or some other audible advantage, compared to the standard 80Hz crossover.
But, concern that we are wasting the capabilities of our speakers should not be a strong factor in our decision regarding what crossovers to use, as the speakers will always be playing content somewhat below the crossovers we select. It is also helpful to remember that our speakers have to play a very large frequency range, and that relieving them of some low-frequency burden (which requires a disproportionate amount of the amplifier power assigned to that channel) may help them to play their entire frequency range more effectively, and with less potential distortion.
Our subwoofers, on the other hand, have just that one specialized job to perform--playing low frequencies. And, as noted earlier, they typically do that single specialized job much better than even very large tower speakers. Particularly where bass-heavy movies or bass-enhanced music are concerned, crossovers of about 80Hz or higher are usually likely to improve our sound quality. In considering the use of crossovers for 5.1 movies, and for bass-enhanced music, it is important to realize how much additional low-frequency demand may be placed on our speakers. That is especially the case at higher master volume levels.
It is also worth noting that we may be better able to mitigate negative bass influences with our subwoofers than we can with our speakers. We will usually position our front speakers in specific locations on the front wall where they form an equilateral triangle with our listening position, or where they look good aesthetically. As we drop below 120Hz, and especially below 80Hz, room modes may be affecting more and more of our bass response. In some cases, we may be able to position our subwoofer(s) more strategically to sound good just for bass frequencies, where our front speakers have to be positioned to sound good for all frequencies.
If the subwoofers are better positioned with respect to room modes than the front speakers, then transferring more of the bass load (below about 80Hz) to our subwoofers may help to promote a superior frequency response with better sounding bass. This is always an issue that has to be resolved on a room-by-room basis, but it helps to understand some of the reasons behind the typical advice to use 80Hz or higher crossovers.
I think that the more that we understand how much more capable our real subwoofers are compared to the woofers in our speakers, the easier it is to use slightly higher crossovers. It isn't just a matter of woofer diameter, it is also a function of the subwoofer driver's motor strength and excursion capabilities (literally moving in and out to displace air and create SPL), the cabinet volume, the amplifier power, the DSP employed, and even the tuning point of a ported speaker, compared to both sealed and ported subwoofers. The subwoofer is typically far better at playing frequencies below about 80Hz or so, with higher SPL and less distortion.
A final crossover issue which is worth exploring involves bass localization. A subwoofer can be localized when the bass sounds that we hear are obviously coming from the subwoofer itself. Early listening tests seemed to indicate that most people couldn't locate the specific direction of bass sounds below about 80Hz (or a little higher in most cases). At very low-frequencies, bass sounds are definitely not as directional as they are for frequencies in our more normal listening range.
It was for that reason that 80Hz was originally chosen as a recommended crossover point from speakers to subwoofers. This is a fairly complex subject, in its own right. In general, where someone cannot identify bass sounds as coming directly from a subwoofer, an 80Hz or slightly higher crossover may be a very good choice. This issue of bass localization is explored in more detail in Section III-D.
Section III-B: Low Frequency Effects Channel:
There is another setting, associated with our subwoofers, which it is important to mention in the context of this discussion. It is not a crossover, but it does control the content we hear in the LFE channel. The .1 LFE (low-frequency effects) channel is a separate bass channel, played only by subwoofers, as long as there are any subwoofers configured in an audio system. (If there are no subwoofers connected to the system, and turned on in the configuration menu of an AVR, then the front speakers will automatically be set to Large, and LFE content will be routed to those speakers.)
The LFE channel is intended to give audio mixers an opportunity to add more bass SPL to 5.1 audio tracks. As explained in several sections of the Guide, we don't hear bass frequencies as well as we do other frequencies in our normal hearing range, so adding even more low-bass SPL, via a separate channel, can really enhance bass sound effects in movie soundtracks.
[It may be worthwhile to distinguish between two-channel music--stereo-- that has been bass-enhanced electronically, and 5.1 channel music. Some bass-enhanced music can contain very low-bass frequencies, and even bass sine waves, and it can be recorded at higher than normal volumes. The subwoofers will handle the low-frequency content in that music via the normal crossovers. But, only 5.1 music (or of course movies) will have an LFE channel which is dedicated to the subwoofers, and which is programmed to be 10db louder than the regular channels. Two-channel music which has been up-converted to surround sound, via Dolby Pro Logic, or by some other surround mode, is still two-channel music and doesn't have an LFE channel.]
As noted in other sections, the LFE channel has a max SPL of 115db for peaks, compared to 105db for the regular channels. That provides an additional 10db of bass SPL for audio mixers to be able to add to their 5.1 tracks. And, that allows selected low-bass sounds to stand-out more. In addition to the obvious difference of 10db between the regular channels and the LFE channel, however, there are also some differences in bass content.
The bass content in the regular channels ranges from the upper-end of the bass spectrum (500Hz) to the lowest frequencies contained in a recording. So, the range could be from 500Hz down to as low as single digits (~2Hz to 5Hz). The great majority of that bass content would be played by the woofers in the regular channels, with the subwoofers taking over at the crossover from the speakers to the subwoofer(s).
The LFE channel has a much more restricted and concentrated frequency range, and is exclusive to the subwoofers, if there are any in an audio system. Bass content in the LFE channel may still go down to the low single digits, depending on the film, but there is a filter which attenuates the bass content above about 120Hz. But, even beyond the difference established by the restricted frequency range, most real content in the LFE channel appears to be concentrated below about 80Hz. And, I believe that is generally intentional.
It is important to understand that the LFE channel is specifically designed to give more weight to the low-bass (and not quite as much to the mid and higher bass) in scenes where low-frequency effects are deemed appropriate by the sound mixer. So, the LFE channel may not be in very obvious or continuous operation during many movies. In some cases, the LFE channel may be used sparingly, and it may just kick-in with more really low and loud bass, whenever a sound mixer wants to emphasize the sound effect at a particular point in a film. In other movies, with sustained and intense low-bass, the LFE channel may be extremely involved throughout the movie. I believe that is the case in Batman Versus Superman, for instance.
* The LFE channel, which exists only in 5.1 (or higher) movies and music, has it's own setting in our AVR's, called the LPF of LFE. Since the LFE channel is intended to contain bass content up to 120Hz, the typical setting for the LPF is 120Hz. And, that is the customary setting which most AVR's engage by default. However, a number of audio experts, including Mark Seaton and Roger Dressler (formerly with Dolby Labs and one of the creators of Pro logic II) believe that it can make sense to experiment with lower LPF settings.
Some people have suggested that relatively little meaningful bass content is mixed into the LFE channel above about 80Hz, as the LFE channel is primarily intended to emphasize lower bass sounds and special effects. That may or may not be correct in general, although some film mixers have indicated that most of their meaningful LFE content is in the lower bass range. It would make sense for that to be the case, since the original Dolby/THX standard was for speakers to crossover to subwoofers at 80Hz, and the <80Hz frequencies were always the ones most commonly associated with subwoofers.
In some cases, setting a lower LPF might emphasize low-bass frequencies a little more, and might also result in slightly clearer bass. Since the LPF is simply a filter, which gradually attenuates volume levels, setting a lower LPF will not completely eliminate bass above the filter, but it will roll-off the higher bass content a little earlier. For instance, an LPF setting of 80Hz would roll-off the 100Hz frequencies by 6dB, and the 120Hz frequencies by 12dB. Doing that would provide relatively more emphasis to the low-bass frequencies, compared to the mid-bass frequencies. That is similar to, but probably more subtle than, approaching the bass from the bottom by lifting the lowest frequencies with a rising house curve.
Mark Seaton has made the point that the more someone is boosting his subwoofer(s), the more that an 80Hz LPF may be helpful in making the bass blend well with the speakers in an audio system. That would particularly be the case where someone was using 80Hz crossovers for the regular channels. Remember that a subwoofer boost lifts all of the bass frequencies symmetrically, in both the regular channels and in the LFE channel. Where significant subwoofer boosts are employed, the bass frequencies above 80Hz in the LFE channel (which are already 10dB louder than the regular channels) might seem to stand out too much in comparison to the lower bass frequencies. Again, that might be more likely to be noticeable where 80Hz crossovers are employed for the regular channels.
Some people may notice a little more bass clarity, and a little greater concentration on the low-bass, with an 80Hz setting. Others may prefer the fuller mid-bass sound with the default 120Hz setting, or may perhaps prefer a compromise setting of 90Hz or 100Hz. The differences among the various settings are probably fairly subtle, depending on the listener, and which setting sounds better is strictly a user preference issue. Although AVR makers typically employ a default LPF setting of 120Hz, there is no absolute right or wrong way to use the LPF of LFE. (FWIW, I do think it is possible that a higher LPF setting might contribute to subwoofer localization, when bass-heavy 5.1 content is playing.)
As a general rule, there may be no particular reason to experiment with the LPF unless a fairly significant independent sub boost is employed--perhaps at least 3 or 4db on top of DEQ, or even more than that without DEQ, and unless crossovers of 80Hz or 90Hz are employed for the regular channels. Or, unless someone is specifically looking for greater clarity in the mid-bass range. There is some additional discussion of methods to achieve mid-bass clarity in the next section on Cascading Crossovers.
[Some additional discussion of the LPF of LFE, including comments from Mark Seaton and Roger Dressler, can be found in the Audyssey FAQ, linked below. It should be noted, however, that one summary comment by another AVS member, at the very end of that discussion, is not correct. LFE material is not "brick wall" filtered at 120Hz. As noted above, the LPF (at any setting) simply rolls-off content gradually, just as any other low-pass filter does.]
"Official" Audyssey thread (FAQ in post #51779)
[It should also be noted, that just as some AVR's only offer crossovers which are already fixed at 80Hz, some AVR's do not allow LFE adjustments. Yamaha AVR's, for instance, do not have variable LPF of LFE settings. In that case, the default setting will be 120Hz.]
Section III-C: Cascading Crossovers:
The concept of using cascading crossovers to increase mid-bass clarity, and to increase dialogue intelligibility, is one that has been around for a while. It may be especially helpful where someone is using significant subwoofer boosts in order to emphasize low-bass frequencies, or to emphasize mid-bass chest punch.
The process is typically defined as setting two crossovers in different places, such as in your AVR and in your subwoofer, to combine at the same frequency. In this case, we won't actually be setting a "crossover" in the subwoofer, although it is often labelled as that on the subwoofer amplifier. We will just be setting a low-pass filter in the subwoofer which corresponds to the crossovers in our AVR. (As explained earlier in Section III, a low-pass filter "passes" frequencies below the set point. By doing that, it regulates the frequencies which a subwoofer is allowed to play.)
As with all setting options, cascading crossovers is something which is implemented after an audio system is calibrated. So, the settings described below are changed from the default settings after the auto-calibration routine is performed.
There are three components to cascading crossovers. First, there are the crossovers from the speakers to the subwoofers, which are typically set at about 80Hz. Surround and height channels may have higher crossovers than that, and that is usually fine. It is mainly the front soundstage, which carries most of the meaningful content and all of the dialogue, that we are trying to affect.
Second, there is the LPF of LFE, which controls a separate bass channel (the .1 low-frequency effects channel) as explained in the previous subsection. Only the subwoofers play the LFE content, and that additional bass content is only present with 5.1 movies and 5.1 music. To implement cascading crossovers, both bass sources in the AVR would be set to the same ~80Hz frequency. So, the LPF of LFE in the AVR would also be changed to 80Hz.
(Some AVR's, such as Yamaha AVR's, don't allow the LPF of LFE to be changed. It always remains at the default setting of 120Hz. If so, it is no problem. Setting the LPF in the subwoofer itself to 80Hz will still have full effect on the crossovers from the speakers to the subwoofer, and the low-frequency effects channel will still roll-off a little faster, too. So, the concept of cascading crossovers will still work.)
The third component is the low-pass filter (LPF) in the subwoofers themselves. As noted above, that filter may be labelled as a "crossover" on the subwoofer's plate amp. It controls how high in frequency the subwoofer is allowed to play before starting to roll-off. To make the two bass sources in the AVR cascade, it would also be necessary to set the subwoofer(s) low-pass filter to the same ~80Hz frequency.
That will often be done with an analogue knob on the plate amp. There may be an "On/Off" switch or an "In/Out" switch which allows users to engage their own LPF. If there is such a switch, setting it to "On" or "In" depending on the switch, will enable the analogue knob to control the sub's low-pass filter. Cascading crossovers occur when both bass sources in the AVR approximately correspond to the LPF ("crossover") in the subwoofer.
(If the analogue knob on a subwoofer doesn't allow exact adjustment to 80Hz, just try to get close. Exact correspondence doesn't matter. If the subwoofer started rolling-off faster at 85 or 90Hz, instead of right at 80Hz, the audible result would still be virtually identical.)
The practical effect of using cascading crossovers is to cause the subwoofers to roll-off faster above the selected crossover point. As explained in Subsections III-A and III-B, the crossover which tells the subs where to take over from the speakers is not a brick wall. It contains a low-pass filter which causes the subwoofer to roll-off gently above a specific frequency, such as 80Hz. That roll-off is typically -24db per octave. When we implement cascading crossovers, we are making the subwoofers roll-off faster, above a certain frequency, so that less bass will leak into frequencies above that crossover. As explained later in this section, that bass leakage into frequencies above about 80Hz can sometimes affect male voice clarity, among other things.
* I decided to add a little more detail to this idea of rolling-off subs a little faster, above 80Hz, especially where significant subwoofer boosts are employed. Let's look at the LFE channel first and assume an LPF setting of 120Hz. For the 8-note octave between 120Hz and 240Hz, the low-pass filter will roll-off the subs by 24dB. That sounds like a lot, but what that means in practical terms is that the subwoofer will gradually lose about -21dB between 120Hz and 240Hz.
Now, let's assume that someone wants to use a fairly significant subwoofer boost. An 8dB boost using something such as DEQ, or through independent subwoofer boosts, would not be at all uncommon. Instead of being down by 8 or 9dB at 150Hz, the subwoofer would still be playing that frequency at about the same volume level that it would have been playing, if there hadn't been a boost. And, the LFE channel is already playing 10dB hotter than the regular channels.
It is easy to see that the subwoofer boost, occurring above 120Hz could make the bass in the LFE channel sound a little heavy. It is also easy to see how that boost above 120Hz could make the subwoofers strain a little more, depending on the overall listening/subwoofer volume. Very few subwoofers can play as clearly, with as little strain or compression, at 160Hz as they can at 120Hz. Rolling-off the LFE channel, above 80Hz, can help to alleviate that potential issue.
The same thing happens in the regular channels, but starting at a lower frequency, if an 80Hz crossover is being employed. Just a 6dB boost (which is very modest for some people) would make the 100Hz frequency play about as loudly as the 80Hz frequency would have played without the boost. Again, it is easy to imagine that the extra boost for the center channel (above 80Hz) could make some male voices sound a little more chesty, as bass fundamentals were amplified. And, that in-turn, could make dialogue sound a little thicker and a little less intelligible. Again, rolling-off the subs a little faster, above the crossover, helps to alleviate that issue where it is an audible problem.
In many cases, listeners have found that rolling-off the subwoofers faster improves overall mid-bass clarity, and especially dialogue clarity. It may also, in some cases, concentrate the bass a little more strongly below the crossover. In my opinion, cascading crossovers are most likely to work well, where the three speakers on the front soundstage are reasonably capable of handling frequencies above about 80 or 90Hz.
I personally believe that unless the speakers on the front soundstage can play 80Hz or so with reasonable power and low distortion, we may be better off setting higher crossovers and letting our subs play frequencies higher than 80Hz. And, we also may not want the subs rolling-off any faster at 80Hz. For that reason, very small bookshelf speakers might not be good candidates for cascading crossovers.
[Combining two crossovers may potentially cause some cancellation at that specific frequency (such as at 80Hz) in some cases, although that may or may not be audible if it does happen. FWIW, I believe that cascading crossovers, acting on their own, are not very likely to cause cancellation at the crossover, or to create an audible problem even if a narrow range of cancellation does occur.
However, as noted in an example on the Guide thread, by @bscool, if some measurable cancellation does occur, it can typically be corrected by either adjusting phase or subwoofer distance, as illustrated at the end of this subsection. In any event, cascading crossovers will typically increase the strength and clarity of the mid-bass frequencies as a whole. And, many other listeners who have tried the process have reported an overall improvement in sound quality and in mid-bass impact.]
Some time ago, I decided to experiment with the concept of cascading crossovers in my system, and I liked the results very much. I will explain what I did and what I liked about the way it influenced my sound quality. (I will note at the outset, that for anyone using Audyssey or some other form of room EQ, nothing about implementing cascading crossovers interferes with the filters set by automated room EQ. This is strictly a post-calibration tweak.)
First, I should explain that I have very capable speakers in my 7.1 system, and I never use crossovers higher than 80Hz. (I think that having reasonably capable speakers, which can handle crossovers below about 100Hz may be a prerequisite for successfully implementing cascading crossovers.) Second, I also prefer to use an LPF of LFE setting of 80Hz. I get better bass clarity when I set my LPF to 80Hz, rather than to the default 120Hz. Potential advantages to using the lower LPF are briefly described at the end of the previous subsection, and in greater detail in the Audyssey FAQ. The fact that I was already using a lower than typical LPF of LFE made me think that I might be a good candidate to try cascading crossovers.
Since I was already getting a very smooth transition at 80Hz, from my speakers to my subwoofers, I decided not to set my subwoofers' internal low-pass filters to that same 80Hz. Instead, I chose 100Hz, for my initial experiments, and then later tried 90Hz and 80Hz. Before attempting to explain what I experienced when I tried this, I should explain the physical mechanism involved. As discussed earlier, when we set a crossover for our speakers, in our AVR's, the speakers typically roll-off below the crossover at 12dB per octave, and the subwoofers roll-off above the crossover at 24dB per octave.
As noted earlier, in theory, the speakers will already be playing -3dB at 80Hz, and the subwoofers will be playing +3dB. But, the subwoofers are still playing the content above 120Hz, although at a reduced volume level, and their SPL still contributes to the overall sound and consumes some headroom from the subs. When, I set a 24db per octave, 100Hz LPF in my subwoofers themselves, I didn't affect the frequencies below 100Hz. But, I increased the magnitude of the roll-off occurring above 100Hz.
What I found when I tried this was that my mid-bass frequencies (up to 100Hz) seemed relatively louder than they had been, and my overall bass clarity improved. I especially noticed that I didn't have to boost my center channel as much as I had been doing, in order to hear clear dialogue. I think this is due to two factors. First, the higher bass content that had been played by my subwoofers was making the front speakers and surrounds a little heavy-sounding in proportion to the somewhat smaller center channel. And, second, since I was already using a heavy subwoofer boost, cutting-off the subs a little quicker imparted less bass coloration to the voices coming from the CC.
This is one of the reasons that I personally prefer not to use DEQ. I don't like boosting the bass in the center channel, with the voice coloration that I notice when I do that. Deep male voices typically only go down to a fundamental frequency of about 90Hz, so bass boosts above that frequency may make men's voices sound unnaturally thick and chesty to some people. As noted in other sections, however, whether we notice that sort of thing, or care about it, is strictly a YMMV issue. (I make up for not using DEQ by implementing a much more substantial subwoofer boost for movies.)
** I also decided to add a little more detail to the explanation of why we may hear more mid-bass and overall clarity when bass boosts don't go above about 80 or 90Hz. Using voices is an excellent way to describe what I think is happening, and that is where I personally notice the additional clarity the most. The human voice is an instrument with a large frequency range. I said that bass boosts above 80 or 90Hz may potentially make male voices sound "chesty". In vocal music, a chest sound is deeper and more resonant than a head tone, which is produced higher in the voice box. The chest tone requires more air, and it resonates lower in the voice box than the head tone does, but it can also sound "throatier", and it has less clarity or "brilliance".
Some consonants, such as "B", "C", "D", "G", "T", "V", and "Z" which all share the same long "ee" sound, may be more difficult to distinguish if they are pronounced with too much chest tone. Some vowels can also be harder to distinguish if more bass sound is added to them, because the voice will sound slightly thicker. I believe that is especially the case if the person speaking has a strong accent, or if he fails to articulate clearly, or if ambient noises in the soundtrack make voices harder to hear clearly to start with.
(When someone articulates, he says each syllable of a word clearly and distinctly. James Earl Jones is a great example of a person with a very deep and resonant voice who is nevertheless very easy to understand. But, he had a speech impediment as a child and worked very hard to learn to speak slowly and with excellent articulation. Most actors do not have that style of speech and that kind of articulate diction.)
Remember also that if subwoofers are strongly boosted, with the normal 80Hz crossover in the AVR, they are only rolling-off at 24db per octave above 80Hz. So, at 100Hz, the subwoofer has only rolled-off by 6db and can still provide quite a lot of bass coloration to male voices. To me, that can make the voices sound a little unnatural as well as more difficult to understand. So, where I may not mind a little additional bass resonance in some music (the cello or the kettle drum, for instance), I may not like it quite as much for some other instruments. And, where I absolutely want it for the low-bass special effects in movies (well below an 80Hz crossover), I may not want that extra resonance at all where the human voice is concerned.
I found that as I implemented cascading crossovers at 100Hz, and then at 90Hz, and finally at 80Hz, I was able to concentrate a little more bass below 100Hz, and then below 90Hz, and then below 80Hz. And, not only did the mid-bass clarity improve with each attempt, but my mid-bass tactile response also increased as a result. That chest punch sensation is explained in detail in Section VII, but briefly, most people seem to feel the sensation most strongly between about 50Hz and 100Hz.
There is some evidence that the sensation may peak for most of us at around 63Hz. That specific number was the conclusion of one study I read years ago, and some subwoofer makers, such as SVS, provide the capability to add a pre-programmed peak at that frequency into their higher-end subwoofer models which have advanced PEQ. If we make our subwoofers roll-off more quickly above 80Hz, by implementing a 48dB per octave filter, we are doubling the roll-off.
So, although there is still some transition between speakers and subwoofers, the subwoofers have rolled-off a good deal more at 100Hz, and they have rolled-off by about an extra 12dB at 120Hz. It is easy to understand how larger subwoofer boosts would allow us to benefit from a quicker roll-off above our selected crossover. And, it is easy to understand how we might be increasing the punchiness of the bass in the range where most people feel those chest punch sensations most strongly.
I offer this method of cascading crossovers as a means of potentially obtaining additional mid-bass SPL and chest punch, combined with potential improvements in overall bass clarity. (The clarity was the real key for me, but again, I use a lot of subwoofer boost for movies.) Determining where to set the LPF in the subwoofers themselves, and what slope to select if that is an option, is something which may require some individual experimentation. But, in my personal opinion, it may turn-out to be an excellent solution for someone wanting to maximize mid-bass SPL and clarity. The setting procedure is summarized as follows:
Setting Procedure:
To recap the procedure to follow in setting cascading crossovers, the following three steps would be performed after a calibration:
First, all of the speaker crossovers in the front soundstage would need to be set not higher than about 100Hz, in the AVR, and 80Hz or 90Hz would be better. So, we would need to have fairly capable speakers for at least our three speakers on the front soundstage. (It might not matter quite as much for surrounds, rear surrounds, height speakers, and so on, as it would be for the channels which carry so much of the fundamental content of both movies and music.)
Second, we would ideally need to be able to implement an LPF of LFE which approximately matches our speaker crossovers. Let's just say we are using 80Hz to make things simple. (If we couldn't adjust the LPF--as is the case in some Yamaha AVR's--we could still try cascading crossovers anyway, as explained earlier.)
Third, we would need to be able to implement a low-pass filter in the subwoofers themselves, or with a miniDSP, which would approximately match the crossovers to our speakers. (Most subwoofers will have either digital DSP, or an analogue knob--sometimes labelled Crossover or LFE--which will enable us to set a low-pass filter for the sub. Some subwoofers may also enable us to manage the slope of that filter. In my case, I rolled-off the bass above 80Hz at 24dB per octave.)
When we implement all three of those settings to coincide: the speaker crossovers, and the LPF of LFE in the AVR; and the LPF in our subwoofer(s), the subwoofers will roll-off much faster above our target frequency of let's say 80Hz, and the mid-bass SPL and tactile sensations will be more concentrated below that frequency. I think it would be generally preferable to make the three low-pass filters correspond with each other, but they don't have to correspond exactly (especially with an analogue knob). And, if the LPF of LFE in the AVR can't be adjusted, there is still some benefit to making the speakers' crossovers, and the low-pass filter in the subwoofer(s), correspond.
*** After trying the 100Hz LPF in my subs for a few days, I experimented with dropping the LPF in the subwoofers from 100Hz to 90Hz, and the results were even better. In my particular case, the bass frequencies were even more distinct, and the center channel was even clearer. In fact, I was able to reduce the volume on my CC by about 2dB, depending on the movie, and still understand dialogue perfectly well. I like using a large subwoofer boost for the very low-frequencies. I am in a large room on concrete, and it can take a significant subwoofer boost to generate the low-frequency sounds and tactile sensations I like.
But, using a large subwoofer boost also tends to make voices slightly thicker, and a little harder to understand, as explained above. We may get so accustomed to a slight bass coloration in voices, and in other mid-bass sounds, that we may not even notice that we are hearing it any more. At least I didn't. But, when that bass coloration is removed, the resulting sound from all of the speakers is much clearer. And, I can especially hear the difference in the center channel.
After getting used to the 90Hz LPF for a few days, which is how I typically like to test setting changes, I decided to drop the LPF in my subwoofers to 80Hz. At that point, all of my settings aligned at 80Hz. With each change, going from full-range settings in the subs to 100Hz, and then successively down to 90Hz, and then to 80Hz, I liked the results better. The acid test for me was when I watched Battle Los Angeles with the all 80Hz settings. It had been a couple of years since I had seen it, but I remembered how difficult it was to hear some of the dialogue during battle sequences. The mid-bass and low-bass were more impactful than ever, and the dialogue was easily understandable at lower volume levels than it had been before, even when I had boosted the CC.
The cascading crossovers make a noticeable difference to me, and to several dozen others who have reported trying them since I first wrote this. I think that the more subwoofer boost we use, the more that this approach may be helpful for us. This is probably not going to be a good solution for everyone. Our audio systems, and individual preferences, are just too diverse for any single method, or setting, to be successful for everyone. But, I definitely recommend trying it, if someone believes that he may be a good candidate, and if his subwoofers allow him to set a lower LPF, which he believes might potentially correspond well with the crossovers he is using in his AVR.
(Just to reiterate, none of the changes I am suggesting will interfere with the automated room EQ that you may be employing, and any of these changes are implemented after a calibration. You would always want room EQ to measure the full capabilities of your speakers and subwoofers. It is only after room EQ has set filters for the various channels that you would implement any limitations to the upper range of the subwoofers.)
Dealing with Cancellation:
I decided to put a brief description of the cancellation that people may sometimes encounter when they incorporate subwoofers into an audio system. Due to a variety of factors, some frequencies may cancel each other, eliminating bass at those frequencies. Cancellation can occur with a single subwoofer interacting with a room, or with dual subwoofers interacting with each other, or with subwoofers and speakers interacting at a crossover. That cancellation may occur either with or without cascading crossovers.
(A separate discussion of dealing with phase cancellation, between the subwoofers themselves, is offered in Section IV-B.)
Cancellation always looks bad on a graphed frequency response. (It looks like a a deep V shape in an FR graph.) And, in general, we would prefer not to have any cancellation at any frequency. But, some cancellation may be pretty inevitable in most HT rooms, even where measuring capability and methods of independent EQ are available. Typically, where we have those capabilities, we are trying to move cancellation to another part of the frequency range, where it is less audible, or to another part of the room, where no one is sitting. Perfect bass in every part of a room is extremely difficult to achieve, even where we have four optimally-situated subwoofers.
I think it is worth pointing-out, for people who may be reading this and who do not have measuring capability, that not all cancellation is even audible. That is particularly the case with relatively narrow areas where frequencies are cancelling. For example, let's say that we have some cancellation centered at about 80Hz. That would be a fairly common area to expect some cancellation from speakers interacting (playing the same frequencies) with subwoofers.
We might not be able to hear that cancellation at all unless it covered a wide area. For instance, the octave between 60Hz and 120Hz consists of 8 notes, so each note in that octave covers slightly more than 10Hz. Most sounds we listen to are very complex, consisting of multiple notes (or frequencies, if the sounds aren't musical in nature), and harmonics (overtones) of those sounds. And, our brains are very adept at filling-in missing information. So, a narrow area of cancellation, centered on 80Hz, might not be audible at all.
Wide areas of cancellation, spanning more than 10Hz, might be more audible, as reduced bass in that particular area of the frequency response. But, for someone who doesn't have measuring capabilities, this is not necessarily something to be concerned about with our audio systems (either with, or without cascading crossovers) unless we suspect (either by listening or measuring) that we are losing some significant bass somewhere.
If we do have some reason to believe that we are experiencing cancellation, either via measurements, or from hearing a specific area of the FR where we think there is less bass SPL than there should be, there are a couple of different methods to reduce or eliminate the cancellation occurring at that frequency. One way would be to adjust the phase control on one or more of our subwoofers. Another way would be to adjust the subwoofer distance control in our AVR's. In either case, we could try playing an 80Hz test tone, or using some steady bass content at about that frequency.
Whether to use phase or distance settings probably depends on what kind of subs someone has (and how they are configured); and on what kind of AVR he has. If a single sub has a phase control, adjusting the phase on that subwoofer may be the easiest way to remove cancellation at a crossover. If a subwoofer doesn't have a phase control, adjusting the subwoofer's distance setting in the AVR will work. If someone has two equidistant subs, Y-connected into a single sub out in an AVR, then the use of the AVR's distance control might be easier to use.
If we are using the phase control on a subwoofer, to make it integrate better with speakers, or with another subwoofer for instance, the following information may be helpful. As I understand it, a distance change of 1/2 wavelength corresponds to a 180 degree change in polarity. So for instance, for cancellation occurring at 80Hz (which is a wavelength 14' long), a distance change of about 7' should reverse the phase completely. If a different crossover is used, there are online calculators which make it easy to correlate frequencies with wavelengths.
The real key to remediating cancellation, though, is to adjust the phase or the distance control gradually, measuring as you go, to determine what setting makes the cancellation either disappear (perhaps by moving it to a more remote area of the room), or move to a higher (or lower) frequency where it will have less audible effect. In an extreme case, and where the ability to measure the FR is not available, it might be possible to hear the bass getting stronger with different phase or distance adjustments, while playing some steady bass content.
The second method, the sub distance tweak, is offered compliments of @Alan P. Both the phase change, and the distance tweak, ideally require the use of measuring equipment such as REW. In the absence of more sophisticated measurement abilities, the use of an SPL meter would be helpful. You would just be measuring a relative increase in the volume level, at the crossover, with changes in either phase or distance settings.
It would be more difficult to do the distance tweak, or the phase adjustment just by listening, unless cancellation at the crossover were quite audible, and that would be rare. (Of course, if it's not at all audible to start with, then many people may not want to adjust the phase, or to perform the sub distance tweak, at all.)
Sub Distance Tweak:
1. Measure the center channel and the subs with REW or comparable software (REW HDMI CH3) using a test tone of 80Hz, or whatever corresponds to the crossover.
2. Add to the sub distance setting of one subwoofer (or of both subs equally, if using an AVR which has a single subwoofer distance setting) in 1' increments. (With some AVR's, you must make sure to back-out of the distance setting menu before the new setting will take effect.)
3. Remeasure.
4. Repeat until you get the smoothest transition around the crossover.
5. If using an 80Hz crossover, it would not typically be beneficial to add or subtract more than about 7' of distance. That corresponds to one-half of an 80Hz wavelength, and a change of one-half wavelength would change the phase of a subwoofer by 180 degrees.
6. We normally have to choose between a good transition for the center channel and the subs, or for the front speakers and the subs. If someone is primarily interested in movies, balance the compromise in favor of CC+sub; and for music, measure with the L/R+sub.
It should be noted again that the procedure outlined above may not be necessary unless someone is either sufficiently curious to measure his results, and discovers something specific in the frequency response, or unless someone hears something that leads him to suspect that significant cancellation could be occurring.
Section III-D: Bass Localization:
As noted in previous sections, 80Hz was chosen as a standard crossover, for Dolby/THX purposes, because it was believed that most people wouldn't be able to distinguish bass sounds as specifically coming from a subwoofer, at frequencies around 80Hz, or just a little higher than that. But, was that assumption really correct? Can most of us really not distinguish directionality in sounds below about 80Hz?
I believe that the often repeated statement that we can't hear directionality in bass sounds below about 80Hz is actually not correct. But, in defense of whatever original research was done, and of whatever conclusions were reached, I believe it was always assumed that there would be at least one subwoofer on the front wall, and that people would have difficulty in distinguishing between bass coming from a subwoofer on that front wall, and the speakers on that front wall.
Based on what I have read about how the concept of "Reference" was developed (as the product of consensus) I am not certain that there were any actual listening tests involved in the idea that listeners would not be able to localize bass below about 80Hz. I recall reading that one of the audio experts involved said that there wouldn't be any localization below about 100Hz, and 80Hz was selected to provide an additional safety margin.
But, if there were any actual listening tests involved, then I don't believe that they could have been very comprehensive, because putting a subwoofer on a side wall, instead of on the front wall, would probably have been sufficient to demonstrate that bass frequencies below 80Hz can be localized. The further that the subwoofer were from the speakers, in a larger room, the easier that the bass localization would be.
Over time, I suspect that the original assumption of a subwoofer on the front wall morphed into a generalized belief (or audio myth) that we can't localize bass at all under about 80Hz, or a little higher. I don't know for sure that early researchers made that assumption, of a subwoofer on the front wall, but it's a reasonable hypothesis to explain their conclusion. In any event, I believe that I can easily demonstrate that we can localize bass sounds under 80Hz, and I think that I can explain why, even if we couldn't, 80Hz crossovers would not be a foolproof solution to potential bass localization from a subwoofer.
Let's start with a music example, and to do that, let's talk about a jazz combo. I listen to a lot of jazz, and examples of what I am about to say abound. A typical jazz combo might involve a minimum of three instruments (counting a vocalist as a potential instrument) and would often consist of four or more. I will take a four-instrument combo as the usual one.
In a typical four-instrument combo, there would always be percussion (a drum set), and there would always be a double (or upright) bass. The drums would carry, and sometimes vary the rhythm, and the upright bass would also contribute to the rhythm and would provide a bass counterpoint to the music. Then, there would usually be a piano, and either a vocalist or saxophonist, or whatever. But, I want to concentrate for a moment on the kick drum, and on the upright bass.
Kick drums go down to a low-fundamental of about 50Hz, and an upright (or double) bass, with the standard four strings, can go down to about 41Hz. So, my first question is, when you are listening to jazz, can you tell where in the room the low sound of a kick drum is coming from? If you are using a surround mode especially, or if it is 5-channel content, can you tell that the drum set is in a particular location in your room, during that recording, and hear the kick drum coming from that location?
Can you hear the bassist strumming the strings on the upright bass, even when he goes really low? Frequently in jazz music, the bassist stays very low throughout most of the song. In some cases, that bass sound is just helping to carry the rhythm rather than trying to stand-out distinctively. But, if your eyes are closed, can you still hear it and point to where in the room the low-bass sounds are coming from?
The fourth string of an upright bass has a fundamental frequency of 41Hz. Even if the bassist doesn't make full contact with that string, the lowest note will be around 45 or 50Hz. Can you hear that the lowest bass sounds of the upright bass are coming from a particular spot in the room, and that the placement of that instrument doesn't change for the duration of the song? If you can, then I believe that simple fact demonstrates that we can indeed localize bass sounds below 80Hz.
This is actually a pretty easy observation for me, because I don't use my subwoofers for music listening at all. So, I am not dealing with any crossovers when I listen to music. In fact, I have six large (full-range), widely separated speakers, in a big room, that I use for music. All six speakers face me across about a 30' width, and a 130 degree arc to my front and sides. So, it's easy for me to pinpoint where specific instruments are coming from, when I am playing 5-channel content, and when I am playing stereo content with a surround mode such as PLIIx.
I had actually been enjoying jazz music that way for years before I started to put what I was hearing into the context of bass localization below 80Hz. And then, more time passed before I decided to write about what I believe is a pervasive audio myth--that we can't localize bass below 80Hz. I believe that any of us potentially can localize bass sounds below 80Hz, if our subwoofer doesn't happen to be on the same wall as our speakers. I think that would be even more likely in a larger room, where the speakers and the subwoofer were more widely separated.
I decided to add some non-music related examples to this discussion. Any of us could make a simple test of our ability to distinguish directionality in bass sounds by playing low-frequency test tones through a subwoofer, with the volume turned-down on our other speakers, or turned-up on our subwoofer. If we can hear test tones, at frequencies below 80Hz, distinctly coming from our subwoofer, then we are hearing directionality in bass below 80Hz.
In fact, when Audyssey or some other form of room correction plays the first series of test sweeps through each subwoofer, for purposes of level-matching the subs, can you distinctly hear those sweeps as coming from each individual subwoofer? I have always found that first series of bass sweeps (thumps) to be quite distinctive, as coming from each widely-separated subwoofer, in succession. The pink noise used in those first sweeps is the range from approximately 30Hz to 70Hz. If you hear that pink noise quite distinctly, as a single low-frequency thump coming from a specific subwoofer, then you are hearing directionality in bass below 80Hz.
Now saying that we potentially can notice something is not exactly the same thing as saying that we definitely will notice. And, that is true with bass localization too. After all, the frequencies played by our subwoofers are mixed together with the frequencies played by our speakers. And, most of the content that we listen to is complex content consisting of both fundamental frequencies and harmonics of those frequencies. Some of us may be more likely to notice than others, but complex content might still be a factor in whether we noticed any bass localization.
Our brains are also very good at adapting to familiar circumstances and expectations. That's why two widely separated speakers can create a phantom (stereo) image at the center point of the speakers. And, it's why we can position a center channel below the level of our screens, or large displays, and still hear the voices coming out of the mouths of the characters whose heads are near the top of the screens. We want to hear what we know we are supposed to hear, and what we are accustomed to hearing, and our adaptable brains do the rest.
But, once our subwoofers move away from where our speakers are located, the illusion that the bass sounds below about 80Hz are coming from the speakers becomes harder to maintain. That is why people who position a single sub on a side wall, or on the back wall, or even in one corner of the front wall, may often have trouble with 80Hz crossovers, or sometimes even with 60Hz crossovers. It's because the ability to hear directionality in bass sounds doesn't just magically go away below 80Hz.
As I demonstrated with the jazz example, or as anyone could demonstrate with a live orchestra or a live band performance, it's actually quite easy to localize bass sounds from percussion sections for instance, below 80Hz, and even below 60Hz. Another example would be a parade, where a band marches by and we hear the low thump of the bass drum as it approaches us on our left, passes us to our right, and recedes in the distance. Most of us probably just accept the myth that we can't localize bass sounds below 80Hz, and we don't apply our real world knowledge to the question.
Had I really thought about it much earlier, I would always have known from personal experience with live performances, that we could localize bass below 80Hz. But, even if there were some validity to the idea that we can't localize bass sounds below 80Hz, would an 80Hz crossover always work? I don't think it would. Let's first address this question by talking about subwoofer boosts. Remember that crossovers work by gradually reducing the volume of a sub, above the crossover, while reducing the volume of speakers, below the crossover.
But, what happens when we turn-up the volume of our subwoofers, relative to the volume of our speakers? Does the crossover still work in exactly the way it was intended, or does the subwoofer now play 100Hz and 120Hz content louder than was anticipated in the crossover design? And, if it plays 100Hz and 120Hz content louder than the crossover's roll-off intended it to, can we localize the bass even more easily in those higher frequencies than we could the 80Hz frequency? It makes sense that we would be able to.
I think that could be another reason that cascading crossovers are such a good solution for people wanting to use significant subwoofer boosts. Cascading the crossover helps the 80Hz crossover to operate more in the way it was intended to, in the case of significant subwoofer boosts.
Here's another reason that I think the 80Hz crossover may not entirely prevent bass localization, where a subwoofer is not located on the front wall, with the main speakers in our HT systems. I have always wondered how the LFE channel factors into this question of bass localization. Even if we accept the original premise that the 80Hz crossover, from speakers to subwoofer, prevents bass localization, what about the 100Hz and 120Hz bass in the LFE channel, which plays with 5.1 content.
That bass content is only played by our subwoofers, if our speakers are set to Small. And, the default setting for the LPF of LFE in most AVR's is 120Hz. To make LFE content potentially even more problematical with respect to bass localization, the LFE channel is already playing +10dB louder than the content in the regular channels, irrespective of any subwoofer boost. (Sub boosts would affect the regular channels and the LFE channel equally, so subwoofer boosts would make the LFE content stand-out even more.)
Couldn't we localize a subwoofer, based on the higher volume level content above 80Hz, from the LFE channel? Wouldn't the louder 100Hz and 120Hz content potentially contribute to bass localization, even if the 80Hz crossover did work somewhat for the regular channels? I have actually preferred using an 80Hz LPF of LFE for several years now, and my subwoofers are well-distributed around the room, so that question would be harder for me to test. But, it has always bothered me that we accept the idea that an 80Hz crossover prevents localization for the regular channels, while most of us are still using the default 120Hz LPF of LFE for the LFE channel, which is already playing bass +10dB louder than the regular channels.
I believe that there is one final factor to discuss with respect to potential bass localization, below 80Hz, and that is tactile sensations. I can't speak for anyone else, but for me, bass TR (tactile response) has directionality too. When I hear a sudden percussive bass sound, at the right frequency, and feel a thump in my chest, I am aware of the general direction from which the sensation originates. It may be helpful to remember that a significant study of chest punch determined that the average frequency where most people felt the maximum impact was 63Hz. And, at least two sub makers offer pre-programmed PEQ boosts at that specific frequency.
If we do feel directionality in bass tactile sensations well below 80Hz, then couldn't some of us localize subwoofers which were not on front walls for that reason too? I believe that we could. In fact, I believe that even very low-frequency tactile sensations--such as thudding and rumbling sounds/sensations can also have directionality. They certainly seem to outdoors, when something heavy falls to the ground and we feel the direction of the vibrations. They certainly seem to in my large room, as well, with my widely distributed subs. That is one reason I have four of them now. I didn't really need more than three subs for overall headroom or frequency response, but I always had a directional hole in the low-frequency bass and TR until I added the fourth subwoofer. I think that was partly due to the somewhat challenging geometry in my 6000^3 room, though.
I definitely think that some of us may be more sensitive than others to directionality in both bass sounds and sensations. But, I also think that there are common characteristics which tend to connect us more than they separate us. I believe that, for most of us, it's just a matter of degree as to whether we are somewhat more aware, or less aware, of bass localization.
In any event, I think that it may be past time to challenge the notion that most people can't localize bass at frequencies at or below about 80Hz, and consider the possibility that we potentially will be able localize 80Hz and lower frequencies, when we begin to investigate our subwoofer placements. As we test different subwoofer locations, we will certainly be able to discover whether bass localization is a factor for us.
FWIW, I think there is a reason why most people prefer to have at least one subwoofer on the front soundstage, and it may not be just aesthetics. In smaller rooms especially, where speakers and subs are closer together, a subwoofer on the front wall should help to prevent localization. But, I have seen specific instances where the best subwoofer positions, in terms of the measured frequency response, produced an unacceptable amount of bass localization. In those instances, as in other aspects of audio, individuals just have to pick the compromises which best suit their personal listening preferences.
Section III-E: LFE+Main:
There is a final setting, found in some AVR's (including Denon/Marantz) which some listeners may be tempted to try. It is called LFE+Main, or double-bass. It does literally double the bass, since it allows exactly the same bass content from the front two channels to be played by both the front speakers and by the subwoofers. With this setting, the subwoofers will continue to exclusively play all of the content in the LFE channel, but both they and the front speakers will duplicate the bass content in the regular channels. (Again, as long as a subwoofer is configured in an AVR system, it will always be the only transducer playing LFE content. That doesn't change with the LFE+Main setting.) It is called LFE+Main because the subwoofers still play LFE content, but now they also duplicate the bass content of the main speakers, instead of just handling the bass content that the main speakers don't play.
Some AVR makers added the feature for those HT owners who really didn't want to set their big front speakers to Small, with a crossover, but who still wanted to be able to utilize their subwoofers. But, as explained earlier in this section, the front speakers may not be able to play all of the bass frequencies in movies, or in some kinds of music, nearly as well as the subwoofers can. So, the front speakers may struggle with some low-bass content, causing audible distortion. And, when both speakers and subwoofers try to play the same frequencies, at the same time, the resulting sound quality can suffer in other ways, as well.
When we use LFE+Main, we are not redirecting the bass from our regular channels to our subwoofers as we are with typical bass-management. Instead, we are allowing the front channel content to be played by both the front speakers and the subwoofers. We can still set a crossover from the main speakers to the subwoofers if we wish, and the subwoofers will only duplicate content below that crossover. But, the front speakers and the subwoofers have very different output capabilities, very different frequency responses, and are EQed differently, so what typically results is considerable cancellation at some frequencies, and random peaking at other frequencies. But, the mid-bass frequencies might sound relatively louder, even if there were some resulting loss of clarity in the bass.
[To clarify what may sometimes be a little bit confusing, LFE+Main only operates when three conditions are met. First, the subwoofers must be shown (set) in the Speaker Configuration menu. Second, the front speakers have to be set to Large. Third, a setting of LFE+Main has to be used. That setting is found in the Bass menu, as an alternative to LFE. When LFE+Main is employed, any crossover still works for the front speakers. It is actually just a low-pass filter, at that point, which affects only the subwoofer. But, it is easier to continue to call it a crossover.
That crossover determines the frequency at which the subwoofers begin to duplicate the bass frequencies being played by the front speakers. So, for instance, if an 80Hz crossover is set, the subwoofers will softly start to begin operation an octave above 80Hz (as described in the discussion of crossovers) and will be at full-effect from 80Hz down. If a 60Hz crossover is set, the subwoofer will be in full operation from 60Hz down, and so on. Meanwhile, the front speakers, acting as full-range speakers, will play all of the bass content in the normal speaker channels down to the lowest limit of their individual capabilities. And, the subwoofers will continue to exclusively play all of the LFE content. The LFE part doesn't change as long as there are subwoofers configured in the AVR.]
It may be important to explain a little more about this idea of having transducers with different capabilities, and which are EQed differently, playing the same content. Audyssey and other systems of automated room EQ set filters for each channel independently. So for instance, the specific stereo content from the left front channel is subject to the EQ for that channel, and the specific content from the right front channel is treated the same way. Content below a crossover point for those speakers is EQed for all of the subs as a whole. And, HPF's for the speakers and LPF's for the subwoofers add gradual slopes which help to prevent excessive peaks in the response at bass frequencies. So, as one transducer (the LF, for instance) drops away, the sub picks up the volume.
But, when both speakers and subwoofers play exactly the same content, with front channels and the subwoofer(s), which have been EQed differently, the EQ that has been done can no longer be relied upon to prevent bass peaks and cancellation at random frequencies. In fact, it is fairly likely that the main speakers and subwoofer(s) will cancel each other at some frequencies, if they try to play the same frequencies.
The distortion that usually results from the LFE+Main setting may produce what is often referred to as one-note bass. That same muddy or boomy-sounding bass can often be heard when room EQ is not operating. Bass clarity occurs when every bass frequency can be heard more-or-less distinctly, without some frequencies peaking and other frequencies dipping or cancelling. LFE+Main rarely allows that clarity for individual bass frequencies to be heard.
To summarize the decision of whether or not to use LFE+Main, it may be helpful to compare it to the use of room correction, in general. People who are looking for bass clarity are more likely to find it when they use some system of room EQ. Turning-off room EQ will generally introduce some muddiness to the sound, as bass frequencies randomly cancel, or boominess as frequencies randomly peak. That may very well create the impression that there is more bass, as described in the first section, because we will typically hear some frequencies much more loudly than we will hear others.
That same thing may happen when we engage LFE+Main. It may increase the apparent quantity of the bass in exchange for some quality, in the form of bass clarity. There is nothing wrong with experimenting to determine which setting we prefer, and proceeding accordingly. But, I believe that it may be important to understand the potential trade-offs we are making, and that it may be important to listen objectively to our overall sound quality. Some AVR makers added the LFE+Main feature over the objections of the creators of Audyssey, and of others, who believed that it was contradictory to the fundamental concept of bass-management, and of room EQ as a means to enhance bass clarity.
One of the distinctions that I might personally make would be between the use of our full-range speakers for some types of music, and their use for other types of listening material. As noted earlier in this section, some people may prefer to use just their full-range speakers for music that doesn't have a lot of low-bass content. In that case, turning off the subwoofers in the Speaker configuration menu, and simply setting the front speakers to Large, would enable the listener to enjoy properly EQed bass, played entirely by his front speakers. And, even if someone prefers listening without the use of EQ, more clarity should be achievable with the front channels either playing full-range content by themselves or with the use of properly bass-managed subwoofers.
As a general rule, I think it's fair to say that, if we need to have our subwoofer(s) playing in order to have sufficient low-bass to begin with, then we are usually better off setting our speakers to Small, and bass-managing them with a crossover. Then, if we ever want even more bass, we can just boost our subwoofer volume to get it. That way, we can still benefit from having the right transducers playing the right content, for their specific capabilities, and we can benefit from having properly EQed transducers and from the improved bass clarity which should result. As with all of our settings, however, this is strictly a YMMV issue.
* The Guide continues in the next post, with Sections IV through VIII.
Regards,
Mike