"Official" Audyssey thread (FAQ in post #1) - Page 541

Ummm, he's asking about resonances, which are peaks, right?

Yeah the "nulls" (dips) are pretty minor, while what I'm referring to as "resonances" are indeed peaks. The one at 30Hz is HUGE (+10dB) and can be pretty much predicted given the size/shape of my HT room (which is of the length/height/width to support a standing wave of about 30, 60, 120...Hz).

Not trying to be combative or negative here (I bought the 3808 because Audyssey has a better rep than MCACC, YPAO, etc.) and I wasn't necessarily expecting it to perform miracles with my sub...but I was hoping it might. So my question still stands...am I doing something wrong or is Audyssey unable to tame +10dB 1/8th octave peaks in the sub-100Hz region? While these might be considered extremely narrow peaks they are the classic reason why subs can sometimes sound terrible (I know - before I bought my PEQ my sub sounded like bigfoot farting in the corner...).

Actually, both nulls and peaks are from "resonances" with the former being from destructive interference of those resonances and the latter being from constructive interference. But you are right; I keyed on the "weird dips" comment and should have grokked the context the OP was using and referring to the peaks as resonances.

How did you measure the Audyssey results?

Using an Infinity R.A.B.O.S. test kit (it includes an inexpensive dB meter, a CD with a bunch of test tones from 20 to 100Hz, graph paper scaled to the test tones, and a neat scale to measure the width/height of a peak). It's pretty slick and, along with my Behringer PEQ, it definitely enabled me to turn my "bigfoot farting in the corner" Velodyne into a sub that both disappeared and responded smoothly (as confirmed by a measured response of +/-3dB from about 24Hz to 100Hz). I've also confirmed the meter's reasonable accuracy in the past with my PEQ set up as an analyzer using a pretty fancy ($$) mic.

Before I even tested the sub's response using Audyssey I could hear that it wasn't doing the job (on music with decent sub-120Hz information the sub would alternate between the boominess of the 60Hz resonance and the anemic response of higher/lower bass notes - I don't have much music that readily stimulates the 30Hz resonance).

P.S. My concern may be rhetorical, since I did use the "Audyssey Setup Guide" for performing the 3808's setup (the one linked to your post) and read therein that "Narrow peaks or dips in the response below 100 Hz that are 1/3 or 1/6 of an octave wide can be improved—but not eliminated—by Audyssey Mult EQ XT."...but I was still hoping for the better....

That last statement is really only relevant if we are talking about the very low frequency range of a satellite speaker. If you are running the speakers as Small then the subwoofer MultEQ XT filters have much finer resolution as they use several thousand points to shape the response over that limited frequency range.

I am concerned that the measurement method for verification is giving you a false sense of security. I am not familiar with the RABOS kit, but it sounds like it is using sine tones measured at a single spot. Those two things make it nearly impossible to get an accurate impression of a subwoofer response because of the extreme sensitivity to meter placement. As a first step, I would recommend taking 6-8 measurements with RABOS in the same places where you had the Audyssey mic and then taking an average across those measurements.

Having said that, it is also possible that a subwoofer placed in just the right spot can excite a room resonance that rings down with a very long time constant. If you have a chance to read Kal's article in Stereophile you will see that one of the big benefits of performing room correction with a time-domain method is that the room resonances are reduced, but more importantly they are "time-aligned" so one doesn't decay much longer than the others. This is something that simply can't be done with a parametric EQ because that only acts on the magnitude response. Also, a parametric filter that is made too narrow actually causes phase anomalies in the response that can be quite audible.

Do you happen to have any web site addresses where I might be able to read some relatively easy-to-understand information on the concept of time-domain eq'ing as opposed to regular parametric eq's? I am not clear whether or not I understand the general concept. I did some searches, but all the information I find seems to simply refer to it rather than really explaining it.

Is the idea that in certain rooms, specific frequencies cause room resonances, and that the resulting longer decay times of these frequencies result in large volume increases at these frequencies? If so, does "time-domain" EQ correction actually make "time-based" corrections -- i.e. does it make cuts/boosts that actually change over time rather than simply boosting/cutting certain frequencies by a specific amount?

Thanks for clarification,

Larry

Larryy, I assume you've seen and followed the advice I posted on the MartinLogan forum on how to perform room correction setups for large dipole speakers. Correct?

Did any of your mic positions fall outside the bounding region and bubble' I recommend in that thread?

Getting too close to one or both of the fronts will usually result in this miss-read.


More than 2x the distance seems suspicious, but if the bass is well integrated, then leave it.

Will be interesting to see how the Pro mic and software measure that sub.


I leave it off on mine and don't have any issues. But will double check this weekend.


Definitely set the L/R to small at around 60hz crossover. The Quest has the same low-freq driver as the Monolith, and it's not that great below 60Hz. You'll get cleaner mid-bass that way.

The normal' Audyssey EQ curve does roll off the highs. Try Flat' and see if what you are missing is the highs. If so, run that way for Music sources. The AVP remembers Audyssey settings per input AND per feed type (e.g. PCM vs DolbyDigtal)

I do not know of such links (although I am sure they exist) but I can suggest two books:
F. Alton Everest's Master Handbook of Acoustics (not as intimidating as it sounds)
Floyd Toole's Sound Reproduction


All enclosed spaces do something like this and the results include (1) increased volume of a particular frequency, (2) decreased volume of a particular frequency (null) and (3) increased maintenance of that particular frequency for a longer time than for other frequencies (increased decay time). Also, the first two effects will vary from position to positing within the room.


In effect but, more precisely, they use a filter that that has a similar time signature as the "mode" it is attempting to correct. The consequence, ideally, is that the filter's effect cancels the energy at that frequency over the time scale of its decay.

A discussion about RABOS has been going on in this thread: http://www.avsforum.com/avs-vb/showt...569&highlight=
Patrich Hart has some informative comments on the subject.

Thanks for the link - I skimmed through it looking for some information that might be helpful to me (sorry but there's some pretty dense stuff there and taking the time to read everything at this point was a challenge). That thread appears to be focused on the setup of two Infinity subs using both R.A.B.O.S. and Audyssey. The only thing I have in common with this is that I am using the R.A.B.O.S. test kit and Audyssey; otherwise, I only have one Velodyne sub. The problem with Infinity subs (last I checked) is that they only come with a one-band parametric EQ; whereas I needed at least two bands due to my 30 and 60Hz peaks. So I chose to keep my old Velodyne sub, buy an upmteen band PEQ and use the R.A.B.O.S. test kit for use in setting up the PEQ.

Nevertheless, what I was able to conclude from the thread (esp. plhart's post) agrees with my conclusion that Audyssey doesn't seem able to help much with the huge peaks that I have, and that using my PEQ first to get the peaks under control (i.e., close to +/-3dB) then running Audyssey is the best approach. Please tell me if I somehow missed something more important. Either way thanks again!

I can concur with your conclusion. I've tried Audyssey with the RABOS adjustments off, and still experienced the 40Hz 'boominess' that occurs in my room. I then ran Audyssey with the RABOS adjustments on, and the bass for music has the smoothness and punch that I really enjoy without the boom.

Audyssey - Thanks for your response. My sub/system setup efforts and conclusion (i.e., bad-mouthing Audyssey's bass correction abilities - nothing nasty or personal was meant...) are merely preliminary since my initial results weren't looking good. My hope was for some guidance to minimize my remaining efforts so that I don't turn this into an acoustics research project. I am going to continue to try different approaches and test/record as much as I can. For now though here's my thoughts about the below.

All of my satellite speakers are set to small (which I've checked before each test just to be sure), with most using a 60Hz crossover. Since I have a 60Hz resonance I am thinking of raising those to 80Hz (which was the single, all-encompassing LFE crossover point used on my previous AVR). Perhaps this will help.

I don't know what signal shape the R.A.B.O.S. CD uses for each of its test frequency bands, but the instructions indeed only tell you to take the measurements in one place. I would love to take the measurements in 6-8 places, but that would mean graphing 23 data points in each of those 6-8 places (and that's getting too close to a research project for me). Besides, I really only sit in one place in my HT room, which happens to be the place where I am taking the R.A.B.O.S. measurements. It may be giving me a false sense, but it is quite audible to me (i.e., with my previous AVR and PEQ I enjoyed smooth enveloping bass, while the Audyssey auto setup EQ on my new 3808 appeared to produce something far less satisfying). Further, my initial response curve efforts, while maybe not quite as scientific as necessary, bears out what I am hearing. It may seem like I am biasing my observations, but in reality, my goal in purchasing the 3808 was to replace the PEQ with Audyssey so that there's one less signal processor in the audio delivery path...I am still wishing for this but so far it ain't looking good....

Will try to read Kal's article - do you have a link to it? As for my sub's room placement, I tried many positions initially and they all varied from bad to worse (my room is literally quite square and any position seems to excite the resonances determined by the room dimensions). So I chose one of the bad (vs. worse) locations based purely on the "spousal acceptance factor" (it's our living room). With my previous AVR (i.e., without Audyssey or any auto setup) and Behringer PEQ I was able to enjoy nearly-perfect bass response (to my ears) for the last 8 or so years. So, unfortunately, the sub will remain where it is (I recently re-tested the spouse on this issue and it is what it is). So unless I can figure out what I might be able to better in setting up Audyssey to correct my sub's response, I'm going to continue on my current path and report back once I have my PEQ set-up again.

Indeed - thanks for reinforcing my observations (not that I want them to be true, but it's better than feeling like the "odd man out"). What I can't seem to reconcile is the gap between Audyssey's stated EQ capabilities (which are supposedly superior to that which one can tweak out of a PEQ - especially with the higher-end MultEQ XT version in my 3808) vs. what we are seeing. I am not stating that Audyssey's claims are false...I just can't determine what I should be doing differently to prove the claims correct (I'd prefer to retire my PEQ if at all possible...). Oy!

http://www.stereophile.com/musicintheround/

#35 contains his review/comments on Audyssey Pro 3.0
#34 contains comments on the Audyssey Sound Equalizer as does #33
#30 contains his review/comments on Audyssey MulyEQ XT

Kal, you've made and published measurements of Audyssey's performance compared to nothing as well as alternative equalization/correction schemes. Do you have any advice for us on how to fairly and accurately measure and evaluate the technology?

And perhaps you could comment on this exchange on the AS-EQ1 thread?

Thanks in advance!

I look at it as having the 'best of both worlds'. I have the ability to initially tame the single bass peak I have in my room with the RABOS settings, and then have the ability to really fine tune the bass with Audyssey. Audyssey fills in the gaps and smooths the bass response curve producing the sound I enjoy. I wouldn't be without either equalization tool.

Hmm...all of this seems to be implying (saying?) that Audyssey MultEq XT alone will not cure extreme room response issues (i.e., that I may not be crazy to believe the results that I've seen). If so, then I need to continue down the path of first correcting my sub's response via my (manually set) PEQ or some new-fangled outboard auto-EQ device (like the SVS AS-EQ1 or Anti-Mode 8033) then run Audyssey to finish the job, or just get the Audyssey MultEQ Pro hardware/software setup.

I'll probably ride the horse I have for now, but the most compelling choice above (should I choose to get more anal about this) is the MultEQ Pro since it can be used on all speakers in my system, not just the subwoofer. (A bit more information: to make matters worse, many of my other speakers are in-wall--more results of the spousal acceptance factor--and we all know that these are also prone to producing some bad response curves). Anyway other than the cost of entry, (which is what exactly since there seems to be multiple components?) and the learning curve, is MultEQ Pro even available to a humble consumer?

The basics still apply to rooms/systems with Audyssey correction, or any other correction technology for that matter. Speaker position, listener position, acoustical treatments all come before electronic correction. Most of us have little or no control over #1 and #2 and electronic correction can only do so much; do you have any bass traps in your room?

Contact Audyssey, the company, and tell them that you are interested in a Pro Kit. They will point you in the right direction.

No (didn't know about these until now - I just did a quick web search, though I can hardly say I learned everything I need to know). And unless there's some inobtrusive way of adding one I doubt it will happen (the SW is in our living room between the couch and a picture window--I'm one of those poor souls who can't afford a dedicated HT and with a spouse that doesn't want to see any A/V components in the first place--and even I'm not willing to make it look like a recording studio). Besides, my SW sounded great with my old AVR-3300 (once I added the PEQ), so I'm a bit hesitant to conclude that my new 3808 wouldn't allow me to get the same or better results without moving the SW, adding a bass trap, etc. OK so I'm a bit resistant to change....


Thanks - I'm going to finish my PEQ/Audyssey set up efforts, then also measure my full spectrum speaker response (using the analyzer module in my PEQ) to see what Audyssey did for my other speakers (like I said they also have problems--albiet less noticeable, and I've lived with them waiting for the 3808 to come to the rescue...). Will report back (that is unless someone verbally hangs me first).

Yes, thank you, that's an excellent resource. I tailored the specific mic positions to my room and usage, but used positions very much in the spirit of your suggestions and satisfying all the usual constraints. The only advice I didn't follow, and probably should have, just for my edification, was to make a separate measurement very close to one of the fronts, to confirm the polarity.


Nope. I was far enough away from the fronts, within their bounds width-wise, and I took measurements both at ear height and a few inches above and below, within a couple of feet of each other depth-wise, over maybe three feet width-wise. (I concentrated on the two seats my wife and I occupy, plus down and forward a bit because people sometimes sit on the floor in front of those seats.) I kept away from any walls or the couch surface and used all eight measurements.


That's what I've chosen to do, and the SW output does sound genuinely good; better than ever before, without question.


Yes! I want to live with this for just a wee bit longer, but I'm champing at the bit to do the Pro pass.


Thanks. It was on by default, I think. I'm happy to leave it off, if that's all that's needed. But I've experimented with turning it back on without being able to reproduce the problem, so there's still a bit of a mystery here. As Kal suggested I should have just confirmed whether or not Audyssey was the source of the problem by turning it off, but that never occurred to me at the time (I also was watching the show with my wife and didn't want to experiment more than necessary to get things working) and then unthinkingly did the usual and deleted the program when it was over. I'm recording alternate showings of it half a dozen times over the next few days to see if I can possibly reproduce this.

Given that I haven't experienced the audio lag on anything else, I'd just blame the program material. But given that turning Auto Lip Sync off fixed the problem, it makes me think, no, there's something about the audio processing that's at fault. But, then, why isn't it reproducible? So I'm a bit baffled at the moment. But at least it doesn't seem to be a recurring problem.


Yeah, I'll definitely do this. I know it's universally recommended. I just wanted to hear what it sounded like before and after I did that. I'm an inveterate experimenter (as long as it's easy ).


I will definitely try this. I had remembered this but haven't had a chance to test it yet. This seems like the best candidate for an explanation to me.

Oh, given that I did the Audyssey measurements with the polarity inverted by the AVP, and assuming that the polarity really should not have been inverted, is it worth experimenting with another non-Pro pass in which polarity is left un-inverted? (Despite the fact that I can hardly hear any difference with polarity changes, as I mentioned in the original post?) What should inverted polarity on both fronts sound like? Would the most noticeable effect be related to how the fronts and the center interact, perhaps, since the fronts are in phase with each other, but out of phase with the center?

There is controversy as to whether absolute polarity is audible. Easy enough to reverse the polarity and see. I don't think I can hear a difference, but I run in correct polarity just because it's right, at least in my mind. Reversing polarity of your front speakers could change they way they interact with your sub at and near the crossover point, so you might need to rerun Audyssey . . . .

audyssey - not trying to be combative, but any thoughts on this (I really would prefer to learn that it's some type of operator error or misunderstanding on my part rather than make unfounded claims about what Audyssey can and can't do)?

I really think that you are basing too much of your thinking on measurements that are not representative of what is going on. Single position measurements with sinusoidal tones are just not indicative of the actual subwoofer response.

We have seen cases where there is such a big resonance in the room that it decays well beyond the length of the impulse responses measured and therefore is only partially addressed. Without having data from your system, I can't really say if that's the case or not. But, I can tell you that putting in a big parametric cut to "fix" that resonance will have no effect on the ring-down time.

Your earlier thought of moving the crossover point up higher is a good direction to take. That will send more info to the sub and allow the higher resolution MultEQ XT filters in the sub channel to apply their correction.

Also, just because you sit in one spot doesn't mean you should only measure in that spot. The Audyssey algorithm needs information from around the listening area (as shown in the measurement pattern in the guide) in order to understand the spatial distribution of the problems.

Re room resonances, let me see if I have this straight: my ceiling is 8.3 ft high, so the wavelength relevant to the dominant (lowest) freq is twice that, 16.6 ft, right? So this freq is (1130 ft/sec)/(16.6 ft) = 68 Hz. (This is indeed near where I see the strongest resonance in a freq sweep.)

I think the issue is to match the measurement system to the goals of the EQ. This means either REW with averaging of several measurement sites or, taking the easy way out, using XTZ with 3 measurement sites. I do not believe that you have to use the same sites or, even, the same number but you must sample the important region.

I would hope that your first efforts would be to optimize room acoustics and to optimize speaker setup. Audyssey has improved every system/setup that I have tried but better uncorrected setups result in even better corrected ones and poor ones are helped.

Whether you need more than plain vanilla MultEQ XT depends on your systems needs and, of course, your personal needs.

Yes and it will be worse if excited at or listened at the boundaries. So, putting the sub on the floor (where it usually is) will maximize. Also, lying on the floor when listening (there's another thread about the pleasures of this) will also maximize one's perception of it.

Thanks for trying your best, but, not intending to sound sarcastic, I kind of expected some of these answers given the notion that "Without having data from your system, I can't really say if that's the case [a partially addressed resonance] or not."

I can say two things (and I will leave it at that until I do a more thorough set up and testing effort):

1. My sub does sound better with Audyssey than with no EQ (it definitely helped make the LF sound like it wasn't coming from the sub's location). So clearly there is more going on than one can see from the R.A.B.O.S. tests (i.e. there is a time domain effect that Audyssey is working it's magic on).
2. But, unlike the effect of my PEQ, with Audyssey I definitely can hear the 60Hz (and in some cases the 30Hz) resonances booming with certain material (that's why I immediately ran the R.A.B.O.S. measurements before proceeding further).

BTW - I didn't mean to imply that taking Audyssey measurements in more than the place I sit isn't important (I did run it with 8 measurement locations), I just don't have the desire to run the R.A.B.O.S. tests in 8 locations (I may try a couple others just to see what happens).

Wouldn't I love to...you probably didn't see my earlier post stating that this is my living room and I have mostly in-wall speakers. The latter is due to the spousal acceptance factor (and you can guess that that nixes the former too).

Anyway, I can't imagine I would even be posting about this issue if I had good room acoustics and could optimize speaker setup (I am hardly what could be called an "audio enthusiast", just someone who can hear a strong bass resonance and weaknesses in my system's response to human voices). Besides, not to sound sarcastic, my perception (however wrong it might be) is that these AVR-built-in room EQ systems are for people like me that can't build a wonderfully designed HT room/system and need something to easily help make their system sound at least decent to a critical set of ears. Conversely, my brother in-law has a gazzillion dollar audio setup with mostly analog components - he would think it heresy to add any digital manipulation to his system, and I can't disagree given that his listening room looks like a recording studio and his stereo setup sounds way better than anything I could ever hope for from my 7.1 system. Whew (I feel better now)....