AVS Forum banner
Status
Not open for further replies.
1 - 20 of 99 Posts

·
Registered
Joined
·
5,897 Posts
Discussion Starter · #1 ·
Is there a commonly accepted "reference" level for reproducing "real life" sound levels? Dolby has I believe the 105dB reference level. Is this generally accepted as adequate?


Relate this question to published frequency responses for various loudspeakers. -3dB points seem essentially useless to me unless that is accompanied by a reference SPL level at which the measurements were taken. For example, a speaker may measure flat (-3dB) to 30Hz, but if that speaker is excursion limited at 30Hz at a level of 75dB then, well, that would seem to be a critical piece of information. Is it industry practice to use ether a 1Watt or 2.83V input at 1 meter for frequency response measurements? If so, then I guess the speaker's sensitivity will reveal the level at which the speaker measures flat. But... that still doesn't really tell you when the speaker becomes excursion limited for various frequencies, it just tells you the roll-off characteristics.


That being said, if you had your "dream" speaker, but were forced to accept a -3dB point at, say, 30Hz, then what SPL level would you ideally like to see those numbers hold to? 90db? 100dB? 105, 110?
 

·
Registered
Joined
·
174 Posts
I would also like to see numbers like that from the speaker manufacturers.


For me that is the only thing that 'could' make larger 3-way speakers have an advantage over smaller 2-ways - SPL capability on lower frequencies. But that is not always the case depending on the design. So to see the measurements for that would be really interesting.
 

·
Registered
Joined
·
1,304 Posts
This is why I love to play in this part of the forum.

Bigus you are mixing up a batch of speaker specifications. 1. Frequency response. How low and high will that speaker produce sound and how much will it deviate from a perfectly flat response. -3db is the accepted threshold for audibility. If you want to increase the sound level of your system you increase its output by 3db. Same with distortion specifications or deviations from flat. This spec gives you some idea of how the speaker sounds.


2. Sensitivity or efficiency: how loud will a speaker play if 1 watt is fed into it. This is meaningless specification in terms of a speaker's ability to reproduce good sound but it will help you to determine which amplifier you will need to generate a particular sound level in your room. Keep in mind that you need 3db increase in amplifier power to achieve this and this means doubling the power of your amp. So if you want to generate really high sound pressure levels ( like 105db ) you may not want to use a speaker with 88db sensitivity rating.


3. Power response: How much sound the speaker radiates in all directions and deviates from live sound. ( I am actually trying to check this definition so I may be in error.)


4. Power Handling: The question I get most often and in some ways the least important spec: How much power can the speaker accept before the voice coil fails.


Impedance: How hard the speaker will make the amplifier work when driving it.


What you guys both are expressing is not merely that manufacturers provide specifications but an article explaining what specifications mean and how to analyze or understand speaker specifications. I think if you search on the achieves of both Stereophile and Audio you may find such articles.
 

·
Registered
Joined
·
2,674 Posts
1. 105dB at the seats is about right, although you might adjust for personal taste (listening consistantly below reference on movies would lower it, sitting in the front few rows of a concert hall may increase it). Distance to the speakers and Rt60 determine your actual output requirements.


2. Acceptable excursion limited output depends on the application. For most music there isn't much content in the last octave, so speakers which are -6dB @ 20Hz and are excursion limited to 82dB @ 20Hz and 100dB @ 40Hz work great run full-range. See http://sound.westhost.com/linkwitz-transform.htm#power for some measurements (there's no clasical music there though). For home theater, such speakers are unacceptable for full-range use and require a 4th order high pass at ~40Hz or 2nd order at ~80Hz.


3. Most drivers suitable for midrange use in home systems (large pro-sound drivers where off-axis response is less important will suffer less) have excessive (I'd say unacceptable) IM + harmonic distortion when used for bass reproduction at reasonable volumes. A 2-way's response in the last few octaves will get you better integration with a sub-woofer, although trying to use it full-range will muck up the midrange from IM distortion and bass notes won't be accurate sounding.


IOW, graphs showing maximum output vs. frequency and distortion vs. level and frequency may be useful in determining what you want to audition.
 

·
Registered
Joined
·
5,897 Posts
Discussion Starter · #5 ·
Drew,


Thanks, very useful information there. It doesn't surprise me that you kinda knew what I was asking - I've been pouring through SL's dipole models, and one factor he doesn't really get into is just how much output is needed vs. frequency to be acceptable. You can back out what he decided on by examining the drivers and crossovers used for the Pheonix and Orion, but I didn't see anywhere his thoughts on whether this was "minimally acceptable" or "should almost always suffice" or "excessive output levels." My guess is that, being dipoles, it is one of the first two... which one is a critical difference for me in working through my own dipolar design requirements.


I'm thinking flat to 40Hz @ 105dB should satisfy all but the most extreme content. I'll check out those measurements on Rod's site - I hadn't run across them before.


BTW, are you using a separate sub (monopole or dipole) to augment your Orions, or are they completely sufficient ran full range? Music only, or HT as well? Pop, Organ, Symphony...? ;)




audiblesolutions,

Quote:
Bigus you are mixing up a batch of speaker specifications. 1. Frequency response. How low and high will that speaker produce sound and how much will it deviate from a perfectly flat response.
What I am suggesting is that there is a distinct need to "mix up" speaker specifications. IMO, a frequency response isn't very useful in itself... when coupled with the excursion limited SPL vs. frequency of a speaker, then you get a better idea of what the speaker is capable of. After all, a single 4" driver might be flat down to 20Hz... at 50dB.

Quote:
-3db is the accepted threshold for audibility.
I thought experiments had shown that under suitable test conditions, .1dB to .2dB was the lowest consistently identifiable level difference. Of course, that doesn't translate well to real music with a higher noise floor, but it would seem to define pretty well the "threshold for audibility."
 

·
Premium Member
Joined
·
8,710 Posts
Quote:
'm thinking flat to 40Hz @ 105dB should satisfy all but the most extreme content.
That's 105dB at the listening position! The intensity decreases inversely with the square of the distance...and 105dbSPL is blooming loud in the mid range. 80 dBSPL at the speaker is 60 dbSPL at about 12' from the speaker.


The reference level (83 SMPTE ... Dolby & THX use a different level) is 20 db SPL below peak. Although bass can be 120 db SPL due to our relative lack of sensitivity to low bass.


The speaker must also be accurate at low levels to maintain full dynamic range. Flat response, btw, is not well received by we humans...perference is for a roll off in the high and low ends.
 

·
Registered
Joined
·
7,335 Posts
Dennis,


Why did the names Fletcher and Munson pop into my head when I read your response? :D



AudibleSolutions said earlier:

Quote:
1. Frequency response. How low and high will that speaker produce sound and how much will it deviate from a perfectly flat response. -3db is the accepted threshold for audibility. If you want to increase the sound level of your system you increase its output by 3db. Same with distortion specifications or deviations from flat. This spec gives you some idea of how the speaker sounds.
You might want to recheck your textbooks. As a general rule, 1dB is considered the "threshold of audibility". This is an estimate not an absolute... we're more sensitive in some frequency ranges then others (See Dennis' post above mine).

Quote:
2. Sensitivity or efficiency: how loud will a speaker play if 1 watt is fed into it. This is meaningless specification in terms of a speaker's ability to reproduce good sound but it will help you to determine which amplifier you will need to generate a particular sound level in your room. Keep in mind that you need 3db increase in amplifier power to achieve this and this means doubling the power of your amp. So if you want to generate really high sound pressure levels ( like 105db ) you may not want to use a speaker with 88db sensitivity rating.
It's not clear where you're drawing this 3dB figure from unless it's from your errant statment in 1) above about audible differences.


Most of us from time to time make the mistake of using the word efficiency when talking about loudspeakers, which is kind of ironic given that loudspeakers are remarkably inefficient transducers.


Without knowing SPL goals at the seats, it's a bit difficult to say what amplifier power is required, or whether a particular speaker is appropriate or not.


Regards,
 

·
Registered
Joined
·
5,897 Posts
Discussion Starter · #8 ·
Good points (edit - good points DE). You are correct, I'm not aiming (at the moment at least) for 105dB at the listening position, but rather at 1m. It will of course be less at the listening position dependant on distance.


And, in this context, it's not that I'm aiming to make a speaker that is ruler flat down to 40Hz, but rather one that is at least capable of the appropriate peak levels, over the whole target frequency range, for some specified environment. Desired EQ can come later... I'd like a good foundation to work from.


Accuracy at low levels is also a concern. My current research has led me to believe that to satisfy both low level accuracy and high level capability, it is preferrable to go the route of larger Sd and lower Xlin/Xmax, rather than the opposite.


Eh... I know you generally do acoustic space design, not loudspeaker design, but you're obviously seen and heard a helluva lot of various loudspeaker designs... so I'd love to hear any further thoughts you have on this.
 

·
Registered
Joined
·
2,674 Posts
Quote:
Originally posted by Bigus



I'm thinking flat to 40Hz @ 105dB should satisfy all but the most extreme content. I'll check out those measurements on Rod's site - I hadn't run across them before.


BTW, are you using a separate sub (monopole or dipole) to augment your Orions, or are they completely sufficient ran full range? Music only, or HT as well? Pop, Organ, Symphony...? ;)
SL is pretty helpful - you might mail him and inquire where his output requirements came from and how much use his Thors get.


I run my Orions full-range on music - symphonic works subjectively at the level I get in the middle of CU's Mackie Auditorium, rock/techno at high but not extreme levels (about 90dBC average at my seat), and jazz/celtic at realistic levels. I've yet to become interested in organ music - there it would be somewhat relevant to see how often the lowest stops actually get used (just like the open E-string on a bass guitar usually isn't played).


I cross them at @ 80Hz 2nd order high pass, 4th order low pass to a ported sub for home theater use so that excursion remains at safe levels even though I usually listen at Reference -5dB with the ocassional full-volume action flick. Obviously there is a real negative qualitative difference going from music to HT bass in my system. I'm going to try switching to an LR4 XO @ 40Hz (put a Linkwitz Transform in my HT path) and if that fix isn't enough I'll eventually build some sealed sub-woofers.


I played with the numbers on W-frame dipole subs for the last octave (target ~105dB @ 1M, since cursory measurements a while ago showed that room reverberation gave me back what I lost from the 8' to each speaker) and found that if I limited my cabinet size to 2' deep this would require a pair of Tumults on each side with an amplifier delivering 500W into 8 Ohms. Not unreasonable if my music called for it although that's a bit much to spend on left+right subs that are only used on movies where I'm a bit less particular about quality.

Quote:


What I am suggesting is that there is a distinct need to "mix up" speaker specifications. IMO, a frequency response isn't very useful in itself... when coupled with the excursion limited SPL vs. frequency of a speaker, then you get a better idea of what the speaker is capable of.
Definately. Since the spectral content of various music flavors is well-known, you could even have maximum output levels specified as a function of recording type.
 

·
Registered
Joined
·
2,674 Posts
Quote:
Originally posted by audiblesolutions
This is why I love to play in this part of the forum.

Bigus you are mixing up a batch of speaker specifications. 1. Frequency response. How low and high will that speaker produce sound and how much will it deviate from a perfectly flat response. -3db is the accepted threshold for audibility.
1dB on all frequencies is audible with both music (more fullness/detail) and pink noise, but isn't perceived as a difference in loudness. Sensitivity to level differences varies with frequency, with 1db being an accepted average threshold of audibility. Perhaps 3db is the accepted point at which the difference appears to be a change in volume?


.5dB frequency dependant level differences between stereo speakers have a noticeable effect on how stable the horizontal positioning of images remains with varying frequency.


The audibility of substantially lower differences has been demonstrated in blind testing. Blips over a short range are less noticeable than trends over a larger frequency range. Such trends are more noticeable than gain/attenuation applied to all frequencies.
 

·
Registered
Joined
·
5,897 Posts
Discussion Starter · #12 ·
Again, some very useful information... thanks. BTW, a pair of AV15's per side would get you close to the Tumults for significantly less money, though the distortion isn't quite as good. The AV12/15 seem to be well suited to dipole use - very clean and open basket, large flared vent. But four dipole Tumults... ahh, that would be nice. :)
 

·
Registered
Joined
·
177 Posts
A few additions to the specmanship: Although 1dB is in the borderline where a level difference starts to be audible as a level difference, much smaller differences (0.1dB, sometimes even less!) are audible as a quality difference. This only matters in comparing equipment, though.


A loudspeaker specification almost never quoted (because it is so embarrassing) is power compression. For a typical speaker, 5dB increments in power input might result sound output like 80, 85, 90, 94, 98, 102, 103 ... The room has an effect in this, too - at loud levels, the air in room pushes against the speaker, even though the same speaker could perform better in a free space. There is a very real limit of how loud a given system can play.


And speaking about realistic levels: You do understand that Dolby etc. recommendations have nothing to do with that, don't you? The recommendations are only to achieve a nice presentation of a movie. I don't believe anybody can play Top Gun (jet engines at close range) anywhere near realistic levels with their home system. (I grew up at an airforce base, and kind of know about that...) And yet there are movies like Apollo 13 and Armageddon... :)


(Btw, the car loudness competition figures are totally bogus: I fail to see how any speaker could produce sound levels above the liquidation point of air (high peak) or below absolute vacuum (low peak), yet pressure levels higher han those points are frequently quoted in competitions... but I, too, could build a meter outputting that kind of figures quite easily. :wink: )
 

·
Registered
Joined
·
2,351 Posts
To get the best imaging and soundstage your speakers need to be pulled well into the room away from walls. This is not the best place to reproduce deep bass. This is why you will find a lot of people with speakers 5-6' off the back wall and a corner loaded sub.


"typical bass sub satellite combinations" cant get it right. But if you look higher up the scale, monitor/subwoofer systems can truly bring it home.
 

·
Premium Member
Joined
·
8,710 Posts
Randy....


Several possibilities exist:

1. He doesn't know how to properly set up such a system for his room;

2. Orchestra music doesn't follow the same rules of physics as other forms of music;

3. or, I didn't read the article and don't have a clue why he feels that way and can't really comment directly on his opinion in the matter.

Quote:
To get the best imaging and soundstage your speakers need to be pulled well into the room away from walls. This is not the best place to reproduce deep bass. This is why you will find a lot of people with speakers 5-6' off the back wall and a corner loaded sub.
Pulling the LCR's into the room (in the absence of other treatments) is a function of SBIR and, for big, full range speakers, getting them way into the room is necessary. (Again, in general, you'll find the bigger the speaker, the further from the room boundaries.) On the other hand, sub placement to provide smooth bass response at the listening position will almost always require a position, or positions, much different from the front speakers. Putting the sub in the corner can boost the in room SPL of a poor sub (meaning the sub isn't adequately sized for your room), but such placements are also not generally the best for smooth response. High Q subs (peaky, boomy, little devils), like corners.
 

·
Registered
Joined
·
2,627 Posts
Dennis,


I will have to pull the article. I can't recall what his reasoning was other than there is generally a "hole" at the crossover point between the subs and the satellites. Before I misquote, I will check the article, but the point that was not made I am sure in the article is that you could never get smooth bass response from a full range speaker because the bass needs to be someplace else. That is what you are saying or have I misunderstood?


Who was the first to come out with a sub satellite combination? That development was more of a breakthrough than I thought. I thought it was more of a WAF at least at first.
 

·
Registered
Joined
·
5,897 Posts
Discussion Starter · #20 ·
Sounds like the article in question was considering true "satellites" with very limited excursion at 80Hz, and not the much more capable bookshelf or monitor types. I think there's a big difference.
 
1 - 20 of 99 Posts
Status
Not open for further replies.
Top