[youngho]

Howdy,

I recently had another look at Floyd Toole's helpful primer on "Getting the bass right" at http://www.harman.com/wp/index.jsp?articleId=122, and one of the questions that I never was able to resolve is as follows:

As I understand it, reverberations are the summation and persistence of various echoes off room surfaces over time. Different frequencies may have different reverberation times, depending on interactions with the room surfaces. Bass modes are standing waves with frequencies dependent on room dimensions, assuming a rectangular room. Standing waves essentially represent resonance.

Unless I'm mistaken, these resonances should produce longer decay times, as well as amplification of the amplitude of these frequencies. Say, for example, you have a room with a resonance at 56.5 Hz but not at 70 Hz. It seems to me that an impulse response for a 56.5 tone should have a different "envelope" than one for 70 Hz, even after adjustment for differences in level produced by the formation of the 10 ft standing wave (I understand this adjustment to essentially represent equalization). I had thought that equalization would be helpful in terms of reducing the "boominess" of bass peaks but that these resonant frequencies would have still longer decay times compared to non-resonant ones. I had thought that room treatment would be necessary to provide absorption and reduce standing wave formation in order to address the problem of prolonged decay time. Is this wrong?

Page 17 of Toole's primer shows the effect of equalization on a troublesome peak. As he notes, "the long ringing of the original room is transformed into a well damped, tight, transient response." This suggests that I am wrong, but I don't understand understand why. Is it that the differences in decay are much less important than controlling the differences in level due to resonance?

Thanks for any assistance in this matter.

Young-Ho

[Ethan Winer]

Young-Ho,

> As he notes, "the long ringing of the original room is transformed into a well damped, tight, transient response." This suggests that I am wrong, but I don't understand understand why. <

You're not wrong in practice. But it is possible in theory to reduce modal ringing if you can devise a filter that is the precise inverse of the room's inherent filtering. I wasn't there when Floyd Toole did that test, but I'll bet a large sum of money that he had to mess around with the EQ for a very long time to achieve the reduction in ringing shown in that graph. More relevant, if he moved the measuring microphone even an inch I'm sure the improvement would be reduced significantly.

--Ethan

[youngho]

Thanks!

[denverdoc]

Youngho,

Hi again. Maybe I can help some here somewhat--an impulse is a mathematical entity of infinite amplitude and infinestimal duration such that the area is 1.  I believe the mathemitician who first defined and explored its was a guy by the name of Paul Dirac.

The pure beauty of the concept lies in the response of any systems response to it.  This is the transfer function (freq and phase response to any and all inputs). The impulse does not have associated with it a particular frequency as all frequencies (by means of Fourier transform analysis) are contained within.

Sometimes a square wave of a certain frequency is used in a similar manner as it will also contain plenty of info, maybe thats what is being cited here?
John

[youngho]

Ah. I had understood an impulse to be a very loud, very brief sound and the impulse response to be the measurement of what happens after that sound. I had thought that the impulse response would initially show the impulse itself, then direct reflections, then indirect reflections. I had thought resonances in the room would increase the amplitude and prolong the duration of the indirect reflections portion of the impulse response in various ways.

Toole does not actually mention a specific frequency in the "How to get the bass right" paper. Rather, he notes that equalization of the single troublesome peak in that particular room improved the transient response and shows an impulse response diagram.

I had previously wondered whether a very loud and very brief 56.5 Hz tone might produce more and longer indirect reflections than a 70 Hz tone in a room with a 56.5 mode but no 70 Hz mode, but I had not realized that impulses must contain all frequencies. Is there a reason this is so? Is there a reason why a single frequency could not be used as the impulse?

Young-Ho

[denverdoc]

had previously wondered whether a very loud and very brief 56.5 Hz tone might produce more and longer indirect reflections than a 70 Hz tone in a room with a 56.5 mode but no 70 Hz mode, but I had not realized that impulses must contain all frequencies. Is there a reason this is so? Is there a reason why a single frequency could not be used as the impulse?

Youngho,

You are absolutely right in that the impulse response of a speaker in a room will first define the transfer function of the speaker, and after a short delay begin to define the transfer function of the speaker/room interaction in the form of reflections. That is why the speaker analysis throws out the later data. As to your question above, sure exciting a room at a resonant frequency will show up problems that a higher tone will miss. As to why the impulse itself contains all frequencies is simply a mathematical reality, (and if you really want to understand why consult any good engineering mathematics book or signal processing primer) it is not something that one really has a choice over, and of course no electronic device known has the propoperty of infinite slew rate, and zero energy storage which would be necessary to create such a spike. Nor would you necessarily want to throw such a thing at a speaker, so practical measurement techniques use modified impulses, or more recently  employ a sequence of random digital  signals with certain properties.

Point being there are many ways to skin a cat to get the same information. My DEQX does it the old fashioned way with logarithmically swept sine waves that covers the frequency band of interest. But such measurements are still of limited value for defining the room as they pertain only to one point in space with a given speaker placement.  

If you are simply trying to get a read on the room from a particular seating position, PC generated test tones at 1/12 or 1/24th octave spacing works plenty well to ID room modes, assuming you have a calibrated speaker and mic.  If you are simply trying to flatten the bass response for a particular speaker in a given room, then all you need is a calibrated mike. Many guys are using the Behringer pro audio gear to good effect for bass eq. Hope this helps and not further confuses,
John

[youngho]

Thanks for the explanation about impulses! Unfortunately, I do have a few things to read before getting to an engineering mathematics textbook or signal processing primer. Actually, I'd probably have to read a few other textbooks to review basic concepts before even getting to those. It's been too long since I've used any math more advanced than the most simple calculus.

Anyway, in informal terms, it sounds like equalization may help fix "boomy" and "loose" but not necessarily "slow" bass due to modal resonance.

Young-Ho

[Occam]

youngho,

Thanks for the reference to O'Toole article. Its very interesting reading. If you're looking for some web based references covering varying levels of mathematical sophistication, you might look here. Although the references are specifically about dsp, it will cover impulse responss and analysis.
http://www.audiocircle.com/index.php?topic=16717

[denverdoc]

Youngho,

Not sure what is meant by slow bass as it all tavels at the speed of sound. But I think your intuition is right on, parametric EQ can only take you so far, but that is a long ways indeed from the +/- 10 to 20 dB response of your average untreated setup to maybe one with a third or fourth as much. Unfortunately, much easier to tame a peak than fill in a notch, or even better results might accrue.  But once you have done so, you will never want to turn back--everything sounds cleaner, faster, tighter, more articulate, etc.

As Ethan points out digitally implemented filters that apply the inverse transform are more powerful still, and likely in the next 5 to 10 years will become a de facto standard as analog signals will likely exist only at the speaker terminals and all sorts of DSP wizardry will be used to tune the speaker and room in advance of that point.

But in the end, as always much easier to make a sows ear from a silk purse than vice versa.    We will still need excellent drivers, cabinetry, exercise good judgement for speaker location, and at least employ a modicum of room treatment.
John

[youngho]

Occam: Thanks for the reference! I'll have to work my way through those links in the near future.

Denverdoc: I'm never quite sure, either, what "slow" bass is, which why I only used it informally. Perhaps a long decay time for a frequency-specific approximation of an impulse (in the sense that a handclap is another approximation of an impulse) might make something sound "slow," whereas a fast decay time might make it sound "fast."

Just the other day, I was just idly imagining a home theater setup using the upcoming NHT loudspeakers using DEQX DSP room correction and wireless subwoofers. The technology really looks very exciting.

Young-Ho

[denverdoc]

Youngho,

Well as you likely know the NHT is based on DEQX and from all who have heard it, it is a very fine speaker. I have for years resisted the urge to drop 3 or 4 k on a pre-pro for HT, but decided to spend 3k for a DEQX because for one it allows you to assemble xover less speakers, use very steep xo filters and correct the impulse response to something close to perfect. I will easily save this much money on speakers alone by building them from finest drivers I can gey my hands on, as the usual retail is 4 to 5 times what the parts cost. Plus it functions as a digital pre-amp, and best yet you can optimize 4 different systems in the house with it, and with the right cabling could switch between them on the fly. Add this and a lot of other features, reasonably priced if not a bargain.
John

[Ethan Winer]

Doc,

> As Ethan points out digitally implemented filters that apply the inverse transform are more powerful still, and likely in the next 5 to 10 years will become a de facto standard <

DSP may appear more in the future, but it will never be truly effective. Again, the more you correct for one location, the worse the results are somewhere else. Even a few inches away. Especially if the DSP tries to do more than just EQ. Just imagine all the added echoes and ringing the person one seat over will hear.

--Ethan

[denverdoc]

Ethan,
With all due resoect, this is nonsense. The solution you presume via DSP is for one point in space, most are averaging now, and the more forward designs incorporate neural nets to decide on which differences across a region in space are most meaningful and seeks to average these so that one has a more or less uniform listening space that is considerably improved on what existed pre-DSP, and to avoid artifacts as in small differences in phase. After all in any decent set up, one can move their head several inches to feet w/o a huge change in the character of the sound, these designs seek to incorporate whatever logic our brain uses to accomplish the same feat!
Cheers,
John

[John Casler]

I think I have solved those very problems with a predictable "future" technology.  If any one has the dough for the Intellectual Property fees, I will disclose for a 50/50 split (PM only)

[denverdoc]

John.

Thanks but I "borrowed a stereotactic brain surgery vise" that keeps me from squirming too much during crucial playback moments.
John

[John Casler]

John.

Thanks but I "borrowed a stereotactic brain surgery vise" that keeps me from squirming too much during crucial playback moments.
John

I too use that same device (currently) but the future does hold hope

I have seen how it will work and it is truly amazing.  Simple in its concept (obviously since I thought it up) yet extremely complex in its execution

Here's the plan:  

 

[Ethan Winer]

John,

> most are averaging now, and the more forward designs incorporate neural nets to decide on which differences across a region in space are most meaningful and seeks to average these <

They can average all they want, but the problem still remains that whatever correction helps in one place by definition hurts somewhere else. Yes, they can minimize the damage in other locations, but the inevitable tradeoff is to provide less correction in any place.

Just as important, if not more so, is that these DSP solutions cannot cure modal ringing, or lower the Q (broaden the bandwidth) of modal peaks. All they can address is the crude response, but not the time-based problems which are at least as important.

--Ethan

[denverdoc]

Ethan,

I guess we are headed for a semantic mire:

Would like your comments Ethan if you have time re this:

http://www.audioholics.com/techtips/specsformats/AudysseyMultEQ3.php

J

[ctviggen]

My first comment would be to ask what kind of sub they were using, as the sub's response is horrible.  I'd like to see a phase response over frequency to see what they're doing with phase.

[Ethan Winer]

John,

> Would like your comments <

Sure, I have plenty of comments.

This is from the text:

the room had been carefully treated with a mix of moveable and fixed absorptive and diffusive panels

So right off the bat they're starting with acoustic treatment. Though how good that treatment is, or whether the reviewer even understands the various types of treatment, remains unknown. Since he referred to "foam diffusors" I'm not optimistic.

“Standing in the Shadows of Motown” ... I have never heard such a monumental improvement in the sound

I have that DVD and I love it. But the bass on the live tracks, which is what the reviewer is referring to, is very boomy sounding. So I have to conclude that "I have never heard" is the defining portion of that sentence. The rest of the review addresses only imaging and other non-bass issues, where I've been discussing the bass range only.

That said, and not to be a party pooper, I take all reviews with a large grain of salt. Even when they're written by real audio professionals.

Finally, I made all of the points I can make at that same forum a few days ago in THIS thread. The concepts of modal ringing and modal bandwidth were completely lost on those guys. You'll see other examples of ignorance early in the thread, such as Pat commenting on my company's products without even knowing what they are.

--Ethan