Acoustic Weirdness..

[hemiram]

I would love to hear/feel 4 subs, that would be great! You must have a huge room just to put them where they won't be in the way, or are you using them for tables or something?

 

[sploo]

I have almost no passive treatment in my room for bass yet I am achieving +-5dB response from 20Hz to 20kHz at the listening position averaged over 6 seats! This was an impossible feat without having 4 subs in my room.

Wow! How on earth are you managing that?

I'm not talking about the 20-200Hz range, but the 1-20kHz range. My room suffered from huge comb filtering problems, with 20dB+ peaks to nulls. A fair bit of room treatment and experimentation/measurement got it to within 10dB for most of the range.

Do you have very high ceilings and/or walls that are a long way from the speakers?

 

[gene]

Wow! How on earth are you managing that?

I'm not talking about the 20-200Hz range, but the 1-20kHz range. My room suffered from huge comb filtering problems, with 20dB+ peaks to nulls. A fair bit of room treatment and experimentation/measurement got it to within 10dB for most of the range.

Do you have very high ceilings and/or walls that are a long way from the speakers?

The key ingredient is selecting very linear loudspeakers! My room is passively treated (but not over treated) by Auralex. The response I am measuring is from one main speaker at a time for all 4 subs playing. Its an averaged response of all 6 listening positions. Audyssey EQ actually helps me achieve this and surprisingly it does most of its correction below 500Hz and barely does anything to my speakers above those frequencies since they are so linear. In small rooms bass is the dominant issue to deal with which is why a combo of multiple subs, passive treatments and active equalization are a must!

 

[Ethan Winer]

Gene,

> If you flatten bass response in the frequency domain, you reduce the ringing. <

I have yet to see compelling evidence of that. I've heard a lot of folks claim it's true, but every time I ask for evidence they always say, "I'll have to get back to you." If you have any before / after waterfall plots showing EQ, or sub placement, or multiple subs, reducing ringing, I'd love to see them.

However, it is not sufficient to show a single microphone position. I know that with extremely careful adjustment you can (possibly) reduce ringing for an area of one cubic inch. What I have yet to see is ringing being reduced for an area that includes both ears at the same time. Let alone six different seats. And yes, I am absolutely serious!

> If you never heard a correctly implemented multi sub system, you don't know what you're missing. <

If you never heard my HT setup with one top of the line SVS sub in a corner plus 40 bass traps, you don't know what you're missing.

> I have almost no passive treatment in my room for bass yet I am achieving +-5dB response from 20Hz to 20kHz at the listening position averaged over 6 seats! <

Let's see some waterfall plots.

> It is well proven and most installers worth their salt are now implementing multi-sub systems for the better installations. <

I understand fully that two or four subs gives a flatter response over a wider area than only one sub. When someone is not willing to look at a bunch of bass traps, and has a wallet fat enough to accommodate four expensive subs without flinching, those additional subs do help. But it solves only the response half of the equation, and does nothing for ringing which is at least as damaging as the peaks and nulls.

--Ethan

 

[Ethan Winer]

Sploo,

> I'm not talking about the 20-200Hz range, but the 1-20kHz range. <

Indeed. I suspect Gene is looking at results with third octave averaging, which hides all the nasty detail called "reality."

But as I recall Gene has a large room, so that helps by keeping the listeners farther from reflecting boundaries where the comb filter peaks and nulls are strongest.

--Ethan

 

[gene]

Indeed. I suspect Gene is looking at results with third octave averaging, which hides all the nasty detail called "reality."

No the reality is you spend alot of time online promoting the only solution you know/offer - your Real Trap product. My measurements are 1/12th octave averaged over 6 listening positions.

I have yet to see compelling evidence of that. I've heard a lot of folks claim it's true, but every time I ask for evidence they always say, "I'll have to get back to you." If you have any before / after waterfall plots showing EQ, or sub placement, or multiple subs, reducing ringing, I'd love to see them.

Then you haven't looked very far. Dr. Toole has authored plenty of papers on the topic as have we. I recall in the past that you aren't well versed in math. If you were, then you would realize the reciprocal relationship between frequency domain and time domain. If you fix freq/phase issues, you also fix time domain issues. Its called Convolution theorem and it works. There is a reciprocity relation between convolution in the time domain and its counterpart in the frequency domain. That is, convolution in the frequency domain becomes a multiplication in the time (or space) domain. This is sometimes called the "frequency domain convolution theorem."

I offer you this elementary reference which does involve some light math but explains this relationship: http://en.wikipedia.org/wiki/Convolution_theorem

I understand fully that two or four subs gives a flatter response over a wider area than only one sub. When someone is not willing to look at a bunch of bass traps, and has a wallet fat enough to accommodate four expensive subs without flinching, those additional subs do help. But it solves only the response half of the equation, and does nothing for ringing which is at least as damaging as the peaks and nulls.

There is no need to spend big $$$ on subs if you are using 2 or 4. Rather than buying 1 $4k sub, by 2 $2k subs or 4 $1k subs and put them in the right areas of the room with proper equalization. Again your spouting off a misconception about ringing. Fixing frequency response resolves the ringing issue. This is usually a more economical and practical solution than stuffing your room with giant bass traps. Regardless a combo of both solutions usually yields the highest dividens if the customer has the budget and space to accomodate.

If you never heard my HT setup with one top of the line SVS sub in a corner plus 40 bass traps, you don't know what you're missing.

I could only imagine the power wasted into thermal heat for your 40 bass traps. No professional installer serious about performance and aesthetics would rely on this type of solution. There are many tools installers can use which are far more effective and practical such as active equalization, multiple subs, correct positioning of listening area and speakers, and passive treatments. Moderation and implementation of all of these tools is far more effective than nuking your room with absorption. Might as well listen in an anechoic chamber

 

[gene]

Measurement of Audioholics Showcase Room (6 Position AVG)

Attached is an averaged response for the primary listening area (6 seats) for the left front speaker with all 4 subs playing.

This is before I did some more tweaking last night to flatten the bass response in the 20-40Hz range. I will include more comprehensive measurements in an up and coming article I am working on.

 

[Savant]

Gene,

Then you haven't looked very far. Dr. Toole has authored plenty of papers on the topic as have we. I recall in the past that you aren't well versed in math. If you were, then you would realize the reciprocal relationship between frequency domain and time domain. If you fix freq/phase issues, you also fix time domain issues. Its called Convolution theorem and it works.

In many cases, addressing the amplitude (frequency) domain issues does indeed have a positive benefit on the time domain response. Otherwise, bass traps wouldn't work at all. However, modal problems can be non-minimum phase. I.e., they are not consistent in the the amplitude domain when measured spatially in the room. Therefore, they are not consistent in the time domain either. This means active controls are less effective - the eloquent "one cubic inch" issue brought forth above.

That you have used spatial averaging has allowed you to overcome this problem for your room. Spatial averaging - like any statistical process - will begin to reduce the uncertainties. Non-minimum phase problems will tend towards the, oh, I don't know - "mean phase." (???)

 

[Savant]

Perhaps the most important thing I learned is that "Generic 'good' listening room ratios are a myth" for which he makes a compelling argument. However, his point does not seem to apply for a playback system having bass management if the subwoofer is in a corner, as in my living room HT. I never experimented with four subs, one in each corner, but I suspect the same thing applies there too. If all the modes are activated equally, their ratio should indeed still matter.

The "optimum room ratios" produce inconsistent results; several papers in the last 10-20 years have addressed their over(mis)use. Many designers are abandoning them in favor of the better tools that are now available.

Also, FYI, the best placement for four subs, according to the paper Dr. Toole cites, is one sub at the midpoint of each wall:

One subwoofer in each corner also has good low frequency support, but does not perform quite as well as one subwoofer at each wall midpoint, in terms of (the parameters measured).

Some other comments:

If you define "good bass" as a flat response, I agree that multiple subs can help. But that does nothing to reduce modal ringing...

As mentioned above, since the methods used to determine the placements of multiple subwoofers involve spatial averaging, there is less need to focus on the effects of ringing. It "comes out in the wash," so to speak.

I also think his comparison of music in a concert hall in the context of why we don't mind longer reverb decays is flawed. That's fine for opera and symphony concerts, but it's not relevant for pop music or jazz

I think it's safe to say there is not - or should not be - much pop music or jazz played in concert halls. It is rare that a concert hall will be designed for purposes other than the performance of symphonic or operatic music. If a concert hall is being designed with the intention to accommodate something like pop music, the design of the hall must allow for the different acoustical needs through the use of variable acoustics. If jazz is being played in a concert hall without this sort of design allowance, the sound quality will be poor since the room is simply not designed for it.

Likewise, he seems to equate "good" but very late reflections in a concert hall that make the orchestra appear wider than it is, to side wall reflections in a home listening room. This defies my own experience, where absorbing those reflections makes the sound stage wider, rather than the other way around.

It is unclear which you are taking issue with - the fact that side reflections in concert halls are useful (true), or that Dr. Toole makes reasonable arguments suggesting that this concept, to a certain degree, could apply to some small room acoustical environments (also true ).

Were we to measure at two locations six inches apart (ear spacing) and combine the results, I believe that would more closely resemble what we hear.

This is addressed elsewhere in Dr. Toole's work, as well as that of others. I.e., spatial averaging of the amplitude response.

As for binaural measurements, the folks at Syn-Aud-Con have been studying the problem for at least three decades. Pat Brown summarized the latest advancements in a recent Tech Topic for Live Sound! magazine.

 

[Savant]

Also, I've read many interviews with movie mix engineers who confirm that all ambience and directional cues are embedded in the sound tracks. They all seem to agree that direct reflections are not desirable in a room because that just competes with what they're doing.

Were they talking about the audio mix for the DVD? Could you cite one (or more) of the interviews? The reason I ask: There are "engineers" doing audio for film that do not completely understand how their work translates to home theaters. They're accustomed to mixing on a huge soundstange for a huge playback venue. They do not fully appreciate how their work translates to the home theater. I have several DVDs where the film soundtrack was used for the DVD. Not good. Translation is poor no matter what your room is like.

In the cases where a DVD mix is produced, the lack of any standard setup for the production environment further confounds the whole process. (But I won't go down that road in this thread! )

 

[Ethan Winer]

Gene,

> the reality is you spend alot of time online promoting the only solution you know/offer - your Real Trap product. <

There's no need for that sort of personal attack. I haven't once mentioned my company or my products in this thread, nor do I usually in other threads discussing the science of acoustics. This is your forum so you can do what you want, but from my perspective that was an undeserved low blow having nothing to do with the subject at hand.

> My measurements are 1/12th octave averaged over 6 listening positions. <

That's better than third octave, but still is low enough resolution to hide the true extent of the peaks and nulls. Especially the nulls.

> the reciprocal relationship between frequency domain and time domain. <

I understand that very well. But no room is a single pole filter! This is why I ask again and again for proof that ringing can be improved in practice. I'm still waiting. As soon as someone shows evidence that EQ can improve ringing for an area large enough to include both ears at the same time, and not substantially worsen the response or ringing elsewhere, I will change my opinion in a heartbeat.

> 2 $2k subs or 4 $1k subs and put them in the right areas of the room with proper equalization. <

Or much better, buy one $2k sub plus $2k worth of bass traps.

> Again your spouting off a misconception about ringing. <

So please prove me wrong!

> I could only imagine the power wasted into thermal heat for your 40 bass traps. <

That's so wrong I don't even know where to start. But I'll try anyway. This came up recently at the AVS forum when Dennis Erskine said exactly what you're saying. So I posted a comparison of total surface coverage for his approach versus mine, using my own living room with 39 traps as the example. (I added the 40th trap only a few days ago.) It turns out that my solution uses only 2/3 the amount of absorbing surface area. The "professional installer" approach Dennis uses treats the entire front wall plus all of the remaining walls up to ear level. That is a lot more absorption than I have in my living room! More to the point, adding bass traps does not suck volume out of the room. If anything, the general perception is an increase in bass level. I have a $150 Pioneer receiver I bought at Costco that powers only the 2 surrounds, plus 3 Mackie HR624 powered monitors for L/C/R, plus an SVS subwoofer. I can easily play concert and movie DVDs louder than anyone would possibly want without even a hint of strain or distortion.

> No professional installer serious about performance and aesthetics would rely on this type of solution. <

That's exactly what puts me ahead of the curve.

> Attached is an averaged response for the primary listening area (6 seats) for the left front speaker with all 4 subs playing. <

With all due respect that information is useless. What I'd like to see is the response at each of the six seats, to at least 1/24 octave resolution, and of course I'd also need to see waterfall data for all six seats too.

--Ethan

 

[Ethan Winer]

Jeff,

> The "optimum room ratios" produce inconsistent results; several papers in the last 10-20 years have addressed their over(mis)use. <

Agreed, and most people don't have the luxury to change their room dimensions anyway. So we endure, and treat the room as best as possible. The good news is that even really bad rooms can be improved enormously with enough bass traps.

> Also, FYI, the best placement for four subs, according to the paper Dr. Toole cites, is one sub at the midpoint of each wall: <

Good point. I did see that, so please just ignore the "in the corner" part of my comment which wasn't needed or relevant.

> since the methods used to determine the placements of multiple subwoofers involve spatial averaging, there is less need to focus on the effects of ringing. It "comes out in the wash," so to speak. <

I'm still waiting for evidence of this.

> I think it's safe to say there is not - or should not be - much pop music or jazz played in concert halls. <

Yes, but that's not the point. The comparison was made that with music in a concert hall we don't seem to mind longer reverb and decays. That applies only to concert hall type music. With pop music and jazz and closed miked bass instruments, we do mind longer decay times. At least I do.

> If jazz is being played in a concert hall without this sort of design allowance, the sound quality will be poor since the room is simply not designed for it. <

Exactly, and the same applies when playing that genre in a living room. Excess modal ringing is always damaging.

> It is unclear which you are taking issue with - the fact that side reflections in concert halls are useful (true), or that Dr. Toole makes reasonable arguments suggesting that this concept, to a certain degree, could apply to some small room acoustical environments (also true ). <

I take issue with the notion that localization cues should be added by the listening room rather than what the mix engineer intended. A side issue, or maybe it's the main issue, is that early reflections usually makes the image smaller, and makes the sound seem to come from the speakers rather than from a phantom sound stage.

--Ethan

 

[Audioholics]

What I'd like to see is the response at each of the six seats, to at least 1/24 octave resolution

Just for kicks, do you have this information available for your room?

 

[Savant]

> My measurements are 1/12th octave averaged over 6 listening positions. <

That's better than third octave, but still is low enough resolution to hide the true extent of the peaks and nulls. Especially the nulls.

A low* frequency resolution certainly provides a high level of detail for the evaluation of peaks and nulls in the amplitude (frequency) response. However, 1/12th octave resolution provides between roughly 2 Hz and 8 Hz resolution between the 21 and 200 Hz bands, respectively. Since modal bandwidth rarely exceeds these sorts of values, there is good scientific evidence that 1/12th octave resolution is "good enough" for the sort of analyses subject to discussion here. When using a 1/12th octave band analyzer, I am often able to pinpoint important areas of concern.

Additionally, the program material must be considered to a certain degree. If movie soundtracks are considered to be the main focus here, it is less likely that the low frequency content will be discrete tones and more likely that it will consist of wider-band content like explosions, rumbles, drones, etc. While I would not advocate an inferior analysis due to program content, this sort of focus does provide for leeway when considering whether to use 5 Hz or 0.5 Hz resolution for amplitude response analyses.

I'm not saying that more resolution couldn't be useful. But 1/12th octave band amplitude response curves should not be summarily dismissed as useless by any means, IMO.

With all due respect that information is useless. What I'd like to see is the response at each of the six seats, to at least 1/24 octave resolution, and of course I'd also need to see waterfall data for all six seats too.

It seems somewhat contradictory that you discredit single-point-in-space measurements, but then request to see exactly that for Gene's room. ( ) I mean no offense or disrespect, but wouldn't this simply provide the response for six points in space rather than one? IMO, this would be six times more useless information according to your own requirements. (And mine, fwiw.)

Spatial averaging is a proven technique, used by many to better quantify the room response and correlate it with what is being heard.

*To clarify, when talking about frequency resolution, "low" is better than "high." Technically, I believe Ethan meant to write "high enough resolution to hide..." in his post above.​

 

[Savant]

> since the methods used to determine the placements of multiple subwoofers involve spatial averaging, there is less need to focus on the effects of ringing. It "comes out in the wash," so to speak. <

I'm still waiting for evidence of this.

I would suggest references [63]-[65] and [70]-[71] from the Toole paper discussed earlier in this thread.

 

[sploo]

Just for kicks, do you have this information available for your room?

I would be very interested in seeing such plots, including 'per chair' readings from gene. Always something to learn...

If it takes one to start it off, then here you go: http://spikyfish.com/acoustics/

I suspect, in comparision, that'll massage a few egos here, as it's the amateur results of an amateur building his own amateur panels .

It seems somewhat contradictory that you discredit single-point-in-space measurements, but then request to see exactly that for Gene's room. ( ) I mean no offense or disrespect, but wouldn't this simply provide the response for six points in space rather than one? IMO, this would be six times more useless information according to your own requirements. (And mine, fwiw.)

Is the issue here that we do hear a form of average (due to having two ears) but that the six seating points in the room will obviously be much further apart? I suppose what you really want is pairs of readings (a few inches apart) for each seating position, but I don't know how you'd make good sense of the readings (i.e. how to combine a pair to form 'what you really hear' at different seating positions).

I was wondering, when gene talked about his results being averages and/or 1/12th octaves, if that would be hiding some detail (comb filtering etc). I'm hoping that is the case, because I'm extremely jealous of the results he's getting .

In general, please keep this thread going guys (as long as none of you want to start hitting one another with subs or traps) - it's a good read.

 

[Ethan Winer]

Clint,

> Just for kicks, do you have this information available for your room? <

ROF,L. I guess I should have seen that coming.

I do not have readings for every seat, but there's no reason I couldn't do that. I just bought a high end (DPA) measuring microphone, and I'm itchin' to use it a few more times just for fun. Shall I go first or do we wait for Gene to post? Not that this is a competition!

--Ethan

 

[Ethan Winer]

Jeff,

> 1/12th octave resolution provides between roughly 2 Hz and 8 Hz resolution between the 21 and 200 Hz bands, respectively. Since modal bandwidth rarely exceeds these sorts of values, there is good scientific evidence that 1/12th octave resolution is "good enough" for the sort of analyses subject to discussion here. When using a 1/12th octave band analyzer, I am often able to pinpoint important areas of concern. <

Yes, 1/12 octave is better than 1/3 octave, and maybe you can identify modes well enough. But in the context of this discussion - how bad really are the peaks and nulls - it's not accurate enough. The drawing below shows the exact same ETF measurement expressed as 1/12 octave and high resolution. Both displays use a gate time of 200 milliseconds, and the difference in resolution is obvious. Since ETF uses different scales (log / linear), for clarity I annotated two peak/null pairs showing the frequencies and dB span.

Click HERE to see the graphs.

It's clear that the nulls are much worse than a 1/12 octave display indicates, and the total span from peak to adjacent null, which is probably the main issue, is also much worse than 1/12 octave shows. Look especially at the peak/null pair at 155 and 163 Hz. Those two frequencies are also less than 1/12 octave apart.

> the program material must be considered to a certain degree. <

Absolutely. This is why I often make the point that nulls can be every bit as damaging as peaks, especially when they fall on a frequency that aligns with the key of the music.

> If movie soundtracks are considered to be the main focus here, it is less likely that the low frequency content will be discrete tones** and more likely that it will consist of wider-band content like explosions, rumbles, drones, etc. <

Some of the movies I watch have explosions, but they all have music.

> 1/12th octave band amplitude response curves should not be summarily dismissed as useless by any means, IMO. <

Agreed, not useless. But not proof of a flat response either, especially when EQ is being used.

> It seems somewhat contradictory that you discredit single-point-in-space measurements <

When did I do that? Heck, I remember when you argued against me about the value of static readings at a single point!

> wouldn't this simply provide the response for six points in space rather than one? <

Of course. But an average of responses is useless because one could be down 30 dB and the other up 30 dB. So it looks like the response is flat at 0 even though both locations are terrible. If I agreed that EQ is useful for low frequency room "correction" then I could see how averaging helps to find one EQ setting that benefits all seats the most. (Or maybe hurts the least is more accurate.) But other than that, which I don't agree with anyway, how is averaging useful when the goal is to know what the room's response really is? Why not just use 1/3 octave averaging and be done with it? Nobody is dissapointed in their room's LF response when it's displayed as 1/3 octaves!

> the center frequencies of 1/12th octave bands very nearly coincide with the notes of the musical scale <

Actually, the standard 1/12 octave frequencies fall almost perfectly between each musical note. I have charts of both here if you'd like me to post some screen caps.

--Ethan

 

[Ethan Winer]

Also:

Were they talking about the audio mix for the DVD? Could you cite one (or more) of the interviews? The reason I ask: There are "engineers" doing audio for film that do not completely understand how their work translates to home theaters.

I haven't forgotten this, and I'm trying to get some answers directly from "the horse's mouth" so to speak. If I'm successful I'll post them here.

> I would suggest references [63]-[65] and [70]-[71] from the Toole paper discussed earlier in this thread. <

I just checked my library and I have none of those. Gee, what a surprise. So do me a favor and tell me the punch line. Do they show the reduction in ringing at multiple locations in an EQ'd room? Can you post a graph or three? Not that I don't trust you or anything...

Thanks Jeff, it's always a pleasure. You make me work hard (>1 hour on today's posts), but in the end I believe we all benefit.

--Ethan

 

[Savant]

> It seems somewhat contradictory that you discredit single-point-in-space measurements <

When did I do that? Heck, I remember when you argued against me about the value of static readings at a single point!

> wouldn't this simply provide the response for six points in space rather than one? <

Of course. But an average of responses is useless because one could be down 30 dB and the other up 30 dB. So it looks like the response is flat at 0 even though both locations are terrible. If I agreed that EQ is useful for low frequency room "correction" then I could see how averaging helps to find one EQ setting that benefits all seats the most. (Or maybe hurts the least is more accurate.) But other than that, which I don't agree with anyway, how is averaging useful when the goal is to know what the room's response really is? Why not just use 1/3 octave averaging and be done with it? Nobody is dissapointed in their room's LF response when it's displayed as 1/3 octaves!

You also wrote in an earlier post above: "As soon as someone shows evidence that EQ can improve ringing for an area large enough to include both ears at the same time, and not substantially worsen the response or ringing elsewhere, I will change my opinion in a heartbeat." I guess I inferred that you wished to understand how different forms of active correction could work for a wider area - not just at a single-point-in-space, or a small area - and not make things worse elsewhere. This is addressed - almost verbatim - in each of the citations from the Toole paper I mentioned above.

Additionally, the averaging methods employed work because they define the minimum-phase problems in the room. I.e., the problems that are consistent throughout the whole room, over a wide listening area. According to the experimental evidence and the subjective results reported in the references you requested, this allows active correction to make noticeable improvements in many situations.

Finally, as I've said numerous times before - and I realize you agree (I'm just reiterating for anyone joining us late ) - active correction is no substitute for a well-treated room. It is simply an additional tool that designers and home theater enthusiasts can employ to address problems with the loudspeaker-room response.