Acoustic Measurements: Understanding Time and Frequency

bblue

Well-Known Member
Apr 26, 2011
360
3
388
San Diego, CA
??? The science of multiple subs says the exact opposite. What the secondary subs do is to cancel out some of the modes created by the other sub. Eliminating modes by definition fixes the problem at that frequency. Here is the actual measurements across multiple seats with one sub from my article on bass optimizations: http://www.madronadigital.com/Library/BassOptimization.html



Now with four subs:



We see clear improvements. The reason is that by eliminating some of the room modes we have actually widened the sweet spot, not made it worse.
Well that certainly is counter-intuitive. What isn't clear to me, then, is what becomes of the room modes? Since they are determined by LWH measurements, shape of the area, etc., how can an additional sub change a fundamental of the room? Or is it a matter of moving it around to a more desirable (less audible) place?

Looking at the 3D spectral plots it looks as if that's what is happening, with the results rather precisely defined for uniformity in a specific area... It would be interesting to see how the plot would look if the corrections were acoustic and not with additional subs? Wider/broader areas of acceptable performance perhaps? Have you compared them?

Bass frequencies radiate in all directions. Not sure what you mean about directional speakers in this regard. How did you find out it was doing nothing?
I didn't mean that the speakers were directional. The angle of the speakers pointing downward from the soffit is very similar to the angle of the whole rear wall trap in the back of the room. It's a full width/height slot trap with the slot side facing forward and tilting back starting at about 2' above the floor. Looking at the blueprints, there's a resonating chamber full width at the base (below the 2' level). It looks like the incoming sound is expected to reflect off the trap's back wall (perpendicular) at the opposite angle downward to its arrival. It would meet stiff fiberglass and reflect backward to the wall again at the opposite angle downward, and finally on down to the resonant chamber where it would be absorbed and effectively disappear.

The slots are alternating 1.5" solid, 9/16" open to trap, 5/8" thick redwood. There's a layer of burlap and 2" solid 703 type fiberglass sheets before entering in the compartment. I only have blueprints, no specs about what each area trap is targeting frequency-wise.

I noted LF standing waves when standing at the rear trap, and measured them as well. If the trap was truly effective I should not have a build up there. Also, adding ceiling high corner bass traps at each corner of the rear of the room made a significant change in response in the listening position.

There are other full height small-slot traps at front sides and middle sides of the room, and a ceiling mounted trap with large chamber behind it (ceiling attic). They all have the usual burlap and 703 fiberglass sheets as first-encountered after the slots. None of these other traps have any LF standing waves in front of them.

And you are saying that with respect to inline electronics that are only in the path to the sub?
Inlines in just the sub channel would not be too signficant unless they were really really poor.

--Bill
 
Last edited:

bblue

Well-Known Member
Apr 26, 2011
360
3
388
San Diego, CA

No, there is a huge difference between a listening room and the room where the recording of live instruments takes place.

I think you’re overlooking a fundamental difference between the two. Both will naturally be influenced by a room with the 70 Hz mode you mentioned. The difference is “music production” vs. “music reproduction.” Since the sound reproduction is “artificially” generated by an electronic system, it can also be equalized electronically to reduce or even eliminate the effect of the 70 Hz mode.

Such is not the case with acoustical instruments like the upright bass or cello you mentioned. Neither has a built-in equalization function. So you’re right - the room acoustics where the recording takes place needs to be fixed in order to accurately record the acoustic instrument. But that has absolutely no relevance to the listening room, as each listening room is different with its own specific set of room modes. As amirm mentioned, electronic de-emphasis of a listening room’s modes will allow for accurate music reproduction from a sound system.

But naturally, that won’t be a help if you invite a musician to play his upright bass or cello in your listening room. :) Make sense?
Not entirely. In the listening room, the speakers become the instument. Behavior of the room can destroy the presentation of the reproduction of the instrument in much the same way as the recording of it insofar as frequency/phase/harmonics response is concerned.

In essence, what's being said is that the playback environment is somehow less demanding to speakers than to an instrument either in a recording room, or the same room as the speakers.

My contention is that while a playback room can be neutralized to a degree with EQ only, for critical listening, it's every bit as significant as the recording environment if you want the most accurate reproduction. Not just a reproduction that would satisfy many/most listeners.

Of course, equalization doesn’t eliminate the room mode. No one pretends that it does. And proper parametric equalization does far more than make the mode “a little less obvious.” If the sound reproduction system is robbed of energy (i.e. exaggerated amplitude) at the offending frequency, the room mode is brought back in line with the decay times the rest of the room is exhibiting.

Take a look at these two waterfall graphs that show the effects of parametric equalization applied to a 41.9 Hz room mode. The second graph includes a parametric filter set at 42 Hz, with the overall level of the signal after equalization raised to match the SPL reading the mode was displaying before being equalized. In other words, 41.9 Hz are at the same SPL in both graphs (~90 dB). We can clearly see that the frequency where the mode was located (41.9 Hz) now displays a significant improvement in rate of decay compared to what it did before equalization, and that the rate of decay is now on-par with the room average. As amirm noted, equalization corrected both the amplitude and time domain issues the mode exhibited.
Right, I'll give you that, but the room average is quite long to start with. Ideally, you'd want to bring that down closer to RT60 standards.

Well, since the source of the room mode is typically the room's dimensions – how do you propose we “fix it at the source?”
With acoustic treatments that minimize the effects of the modes and their siblings. Once that's under control, THEN touch up further with EQ if needed.

I get the impression you don’t have any first-hand experience using parametric EQ to tame low frequency response? I’ve never heard anyone who did complain they didn’t get a dramatic improvement in low frequency sound quality.
Actually, quite a bit. And it does make some improvement. I've just never been satisfied with the results by themselves.

--Bill
 

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal

One point that I am slowly understanding is that once you speak about separate subwoofers with separate independent equalization and amplification we must look at at the problem using a new perspective - there are no simple rules. As far as I can understand the Harman algorithms carry an heuristic optimization and interactively test many solutions, converging to an optimum based on in situ analysis. An individual sub with independent controls can perform tasks that could not be considered easily in classical systems, such as null cancellation.

The multi- subs approach had two phases - the single electronics channel with multiple similar units, relying mainly in positioning and equalization, and the multichannel. Amir is addressing the later phase, which unique solution he refers to is proprietary of Harman. The Gueddes technique can be also considered as part of the later approach. IMHO, unless we know what people are exactly talking about in this thread the debate becomes ambiguous and noisy.

Also IMHO, putting an equalizer such as the DCX2696 in the non-experienced hands of amateurs, without a proper recipe book, is just one step towards the bass abyss, unless his bias expectation is so big that he can successfully rely on his ears. ;)

IMHO bass quality is much more than just achieving a flat response at your listening position. The number of variables you have to control to set such a system is very large, and unless you have the help of an expert you risk a long journey in the dark. I have done a few experiments with equalization and was disappointed with the long term results - I could get better FR in the bass, but overall the sound quality was not better. But I am still trying ...
 

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal
Doesn't the multiple sub approach only work if they are all run in mono?

AFAIK, yes. You must sum the two channels to use it in stereo. But after that point you are free to separate them, most probably using digital signal processing.
 

dallasjustice

Member Sponsor
Apr 12, 2011
2,067
8
0
Dallas, Texas
So, if it's true that one must run all four subs in mono for the Harman technique to work, wouldn't it be best for stereo listening to cross them over at the threshold of bass localization to preserve LF localization (>70 Hz)?
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
...................
So when you turned down the volume you DIDN'T see a decrease in the post-ringing as per Amir's graph? [/I]

Nope. Nothing. Not a change anywhere.

Someone can correct me if needed, pretty sure the impulse graph does not show spl per se, but rather a percentage of full scale deflection or something. As it shows 'comparisons' rather than absolute, that explains why simply lowering the volume still gives the same graph as it were.

Any of those 'swings' we saw ONLY came down or changed with the application of a filter, a mere change in volume did nothing to it at all.

Still need to do a bit more to get a better understanding, but *our* (note I am dragging you in with me in this sinking ship haha) thoughts on the matter-or at least this particular one-hold no water.

Which is kinda ironic given the sinking ship metaphor:D
Yes, Ok, it's a power factor rather than an absolute - I don't think it invalidates the hypothesis or sinks the boat :)
The amplitude of the resonance & post-ringing will of course be reduced when the overall volume is turned down - no doubt about it.
So if you reduce the volume of a particular frequency in isolation, it will also reduce the resonance & post-ringing at this frequency.
My point (& I thought you represented the other side of the coin?) was that this is not a time-domain adjustment - it is simply a downward adjustment of the audibility of the resonance & the post-ringing - let's say it is reduced from 80dB to 65dB as per the original waterfall plots (I think). But let's say that the volume is turned up by 15dB, the resonance will be back to 80dB & the same level of post-ringing will be back also.

Now my question/point is that these room modes may not be as noticeable because they are at a lower volume ratio compared to the overall volume but they are still there & I presume still have an audible effect.
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
Yes, Ok, it's a power factor rather than an absolute - I don't think it invalidates the hypothesis or sinks the boat :)
The amplitude of the resonance & post-ringing will of course be reduced when the overall volume is turned down - no doubt about it.
So if you reduce the volume of a particular frequency in isolation, it will also reduce the resonance & post-ringing at this frequency.
My point (& I thought you represented the other side of the coin?) was that this is not a time-domain adjustment - it is simply a downward adjustment of the audibility of the resonance & the post-ringing - let's say it is reduced from 80dB to 65dB as per the original waterfall plots (I think). But let's say that the volume is turned up by 15dB, the resonance will be back to 80dB & the same level of post-ringing will be back also.

Now my question/point is that these room modes may not be as noticeable because they are at a lower volume ratio compared to the overall volume but they are still there & I presume still have an audible effect.
I am really having a tough time following your reasoning here. Let me try an analog. Let's say I build a DAC that has a distortion spike proportional to source at -70 db relative to 0 dbfs (full scale). If I play -30 db, that distortion goes down to -100 db. Now I modify the circuit so that it actually has a distortion product at -100 db relative to full scale. Are you saying this fix is of no value because turning down the volume in the former case accomplished the same thing? Or invalidated the usefulness of the circuit modification that lowered the distortion? Or that because the distortion is at -100 db and not at minus infinity, the "problem is not fixed?"
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
Let me have another go!

What I'm trying to get at is - in theory is room treatment not a better solution to this resonance & post-ringing as it will suppress both the initial resonance & also each reflection that causes the post-ringing whereas a digital filter just suppresses the resonance?

I also have a general issue with the idea that an FFT tells all that is going on so maybe it's not the best tool for using in analysing sound where music reproduction is concerned. Freq & Time might be orthogonal in terms of FFTs but we have to look at the underlying premise of FFTs & check whether it is applicable to music or just to repeating signals?
 
Last edited:

bblue

Well-Known Member
Apr 26, 2011
360
3
388
San Diego, CA
...
I didn't mean that the speakers were directional. The angle of the speakers pointing downward from the soffit is very similar to the angle of the whole rear wall trap in the back of the room. It's a full width/height slot trap with the slot side facing forward and tilting back starting at about 2' above the floor. Looking at the blueprints, there's a resonating chamber full width at the base (below the 2' level). It looks like the incoming sound is expected to reflect off the trap's back wall (perpendicular) at the opposite angle downward to its arrival. It would meet stiff fiberglass and reflect backward to the wall again at the opposite angle downward, and finally on down to the resonant chamber where it would be absorbed and effectively disappear.

The slots are alternating 1.5" solid, 9/16" open to trap, 5/8" thick redwood. There's a layer of burlap and 2" solid 703 type fiberglass sheets before entering in the compartment. I only have blueprints, no specs about what each area trap is targeting frequency-wise.

I noted LF standing waves when standing at the rear trap, and measured them as well. If the trap was truly effective I should not have a build up there. Also, adding ceiling high corner bass traps at each corner of the rear of the room made a significant change in response in the listening position.

There are other full height small-slot traps at front sides and middle sides of the room, and a ceiling mounted trap with large chamber behind it (ceiling attic). They all have the usual burlap and 703 fiberglass sheets as first-encountered after the slots. None of these other traps have any LF standing waves in front of them.
Though a little OT I thought I'd clarify my attempt at describing the rear wall trap in my studio.

First, I should say that I've never really analyzed what any of these traps are, or what their target ranges are. I (stupidly) assumed that they were as I understood them to be when I spec'd the room originally in 1977. Come to find out, this was a bad assumption to make. I dug out the original blueprints, and other than a few minor liberties taken with the sizing of the room by my construction crew, it's pretty much as the blueprints specified. Within a few inches any direction, at least. Nominally 15.5' wide, 18' front to back, and 8'6" height.

All the traps employed are broad band slot traps; The rear wall that I assumed was a bass trap, isn't. It's a full wall slot trap which according to some formulas I found during a net search, is centered around 1500 Hz. I couldn't find any slot trap layouts that included an attached separate chamber (from the area immediately behind the slotted area), so I'm not entirely sure what it's supposed to do except possibly provide a deeper absorption amount. It is roughly about 18" x 22", at the base of (but not directly behind) the slotted area, right at the floor, full width. I supposed it could be a resonator, or perhaps just an area extension behind the slots. Anyway, the whole thing is not in any shape or form a bass trap.

The side wall traps are active in the 3000-to 5000Hz area. That leaves the ceiling trap which has much greater slot width and thinner+deeper slat depths and opens in to a much larger area, as the bass trap. It's the only one in the room. I can only assume (and remember what that's gotten me so far) that in its design it was to be more functional if the speakers were really soffit mounted. Mine are not soffit mounted, but are on the floor on Sound Anchor stands, too close to both side and rear walls.

Unfortunately, because of the size of my U shaped table/bench surfaces, there's not much I can do about the speaker placement. The assembly is about 1' less wide than the narrowest part of the room, and its front edge is only about 5' rearward of the front wall. It can't be moved any further back without relocating a door.

I'll start another thread with more details and trap construction, for comment. Hopefully someone will spot a way to make some improvements. Or perhaps some of the ideas with a different room layout would be useful to others.

--Bill
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
Let me have another go!
Sorry for the late reply. Wanted to get permission to show you some helpful graphs. Just received the kind OK from Allan.

What I'm trying to get at is - in theory is room treatment not a better solution to this resonance & post-ringing as it will suppress both the initial resonance & also each reflection that causes the post-ringing whereas a digital filter just suppresses the resonance?
The whole point of this article is that suppressing the resonance does both in the case of those caused by minimum phase conditions. The preference for electronic correction is that it is very easy to deploy and takes up no space or mess with your decor. Getting 20 to 30 db correction out of acoustic products alone can be very challenging. Our reference theater has a ton of treatment including the entire front wall that is a proprietary bass trap and you still need EQ and lots of it to get a smooth response.

Back to why fixing the resonance fixes both frequency and time domain, here is a single resonance and its response both in Frequecy+Phase domain and Time:



We see the predictable effect we have talked about. The existence of a peak in frequency (top left), results in ringing of the impulse as time goes by on the right. If this is a minimum phase phenomena, then we can apply an inverse minimum phase correction which applies to both amplitude *and* phase at the same time:



The fact that we are correcting frequency response is obvious from that inverse filter. Less obvious is that the phase correction comes along with it and it also gets flatlined per right hand side with its inverse. Fourier theorem says that signals are fully described using their frequency response and phase so the time domain behavior cannot be any different as it corresponds to this view one for one. But we can confirm that. Here is what is going on in time domain:



We see that inverse ringing is created which helps cancel out the ringing from the original resonance. Let's put both views together now:



We see that we have corrected the frequency response (top left), the phase (bottom left) and time domain (right). It is not easy to escape the nature of these :).

While I was trying to get permission for these slides, Allan wanted me to note that the way to be sure to find minimum phase response is to perform spatial averaging. That is, measure at multiple points and average. In doing so, the non-minimum phase contributions get reduced. So you would want to do this even if you are optimizing for one seat.

I also have a general issue with the idea that an FFT tells all that is going on so maybe it's not the best tool for using in analysing sound where music reproduction is concerned. Freq & Time might be orthogonal in terms of FFTs but we have to look at the underlying premise of FFTs & check whether it is applicable to music or just to repeating signals?
The implementation of correction filters can and most often is in time domain. There is no FFT done and then the bins manipulated although that could be done too. But rather, a FIR, IIR, or deconvolution is performed. The data does need to be digital however and as I explained, I see no issues with that for bass frequencies which is the topic of this thread even if one has a stance against them :).
 

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal
(...) Our reference theater has a ton of treatment including the entire front wall that is a proprietary bass trap and you still need EQ and lots of it to get a smooth response.

Amir,
Since you refer to it, I ask a direct question : considering all you know today and looking retrospectively, do you think you could achieve a similar effect in the bass frequencies using only EQ, even lot of it?
 

microstrip

VIP/Donor
May 30, 2010
20,806
4,698
2,790
Portugal
I came across this article which I thought relevant to the discussion here & the one on room treatments http://www.dagogo.com/View-Article.asp?bShowUnpublished=&hArticle=919&PageOfArticle=1

It is an excellent article - as far as I remember Steve's old listening room was designed by Art Noxon using ASC products and also we have a expert forum on the WBF "Art Noxon Hosts A Discussion Corner On Room Acoustics" with some great contributions of Art.

http://www.whatsbestforum.com/forumdisplay.php?234-Art-Noxon-Hosts-A-Discussion-Corner-On-Room-Acoustics
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
It is an excellent article - as far as I remember Steve's old listening room was designed by Art Noxon using ASC products and also we have a expert forum on the WBF "Art Noxon Hosts A Discussion Corner On Room Acoustics" with some great contributions of Art.

http://www.whatsbestforum.com/forumdisplay.php?234-Art-Noxon-Hosts-A-Discussion-Corner-On-Room-Acoustics
Yes, it is an excellent article. Thanks for pointing out that section of the forum to me - I didn't know Art Noxon had a section here - I'll have a read. Apologies if this article has been discussed before but it seems important enough to have this link repeated.
 

amirm

Banned
Apr 2, 2010
15,813
37
0
Seattle, WA
That was a very good article John. I had not seen it post here and it was very appropriate to the topic here. Along the lines of what he explains about speech, popular music also has a bean which is around 4/second. For this reason, you don't want reverbrations that go too far past 0.25 seconds.
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland
That was a very good article John. I had not seen it post here and it was very appropriate to the topic here. Along the lines of what he explains about speech, popular music also has a bean which is around 4/second. For this reason, you don't want reverbrations that go too far past 0.25 seconds.

One of the most interesting posts by Art in that thread (of many interesting posts) is this one which fleshes out what you posted, Amir
http://www.whatsbestforum.com/showt...lligibility-Test&p=51963&viewfull=1#post51963
When we talk about early and late reflections we need to keep track of the difference between early and late reflections.

A) One strong early reflection creates the comb filter effect, sorta like the coloration of sound you hear when you talk through a tube.

B) Many not-so-strong early reflections have no comb filter coloration effect. However, they do increase the perceived loudness of the direct signal.

C) Intelligibility is not ruined by many early reflections, it is enhanced. Intelligibility is ruined by many late reflections.

D) Intelligibility is ruined because of “sound masking”. There are two types of sound masking, spectral masking and temporal masking.
a) An echo, a strong and very late reflection causes sound masking by upsetting, confusing the tempo of a sound sequence.
b) The best sound masking signal is a set of late reflections that sound just like the direct signal but that are phase and time scrambled, as if it became reverberation.
c) Reverberation is like a noise floor, which also causes sound masking. Reverberation is not a reflection, but a condition of sonic space, where all organization of a sound has been lost. The spectral energy is still in the room but without any sense of direction or timing.
d) Early phase and time distorted reflections, inside the 1/30 second Haas window, are pretty difficult to create. It takes time to capture and rearrange the timing of reflections enough to “scramble” reflections. The time it takes to do this is enough delay that it turns the early reflection into a noisy, sound masking late reflection.
 
Last edited:

NorthStar

Member
Feb 8, 2011
24,305
1,323
435
Vancouver Island, B.C. Canada
I came across this article which I thought relevant to the discussion here & the one on room treatments http://www.dagogo.com/View-Article.asp?bShowUnpublished=&hArticle=919&PageOfArticle=1

It is an excellent article - as far as I remember Steve's old listening room was designed by Art Noxon using ASC products and also we have a expert forum on the WBF "Art Noxon Hosts A Discussion Corner On Room Acoustics" with some great contributions of Art.

http://www.whatsbestforum.com/forumdisplay.php?234-Art-Noxon-Hosts-A-Discussion-Corner-On-Room-Acoustics

-----Just making sure that I got those two great links in my audio archives. :b
 

jkeny

Industry Expert, Member Sponsor
Feb 9, 2012
3,374
42
383
Ireland

About us

  • What’s Best Forum is THE forum for high end audio, product reviews, advice and sharing experiences on the best of everything else. This is THE place where audiophiles and audio companies discuss vintage, contemporary and new audio products, music servers, music streamers, computer audio, digital-to-analog converters, turntables, phono stages, cartridges, reel-to-reel tape machines, speakers, headphones and tube and solid-state amplification. Founded in 2010 What’s Best Forum invites intelligent and courteous people of all interests and backgrounds to describe and discuss the best of everything. From beginners to life-long hobbyists to industry professionals, we enjoy learning about new things and meeting new people, and participating in spirited debates.

Quick Navigation

User Menu

Steve Williams
Site Founder | Site Owner | Administrator
Ron Resnick
Site Co-Owner | Administrator
Julian (The Fixer)
Website Build | Marketing Managersing