equalizer question

[WmAx]

How? I thought raising nulls with a PEQ was 'asking for it' given that the amp would be significantly more taxed, and that the artificially induced additional driver excursion of the speakers could, if not very careful...

D e s t r o y T h e m!

Regards

You can not compensate for true inverse energy nulls with equalization(because you simply increase the positive and negative halves that are colliding, thus equaling the same net effect). But I am referring to peaking. Your FR graphs are pretty low resolution(I would recommend you add a high resolution sine sweep for additional aid/verification), but an appox. 35 Hz peak of substantial magnitude appears to exist in your graphs. A single parameteric band can substantially improve this response characterisitic. To appreciablly effect these very low frequency bands with physical absorbers requires a substantial amount of them that would probably be unacceptable to all but those with dedicated purpose specific listening rooms or audio control rooms.

-Chris

 

[Ethan Winer]

Lungs,

> Do you NOT recommend electronic manipulation of room response at all? <

I have 38 bass traps in my living room home theater and the sound is absolutely fabulous, so I have no use for EQ at all. But I appreciate that most people do not have a wife as loving as mine.

I am not totally opposed to EQ, and in lieu of having the space, or money, or spouse approval, in a larger room it might be better than nothing. But in small rooms EQ is likely to create more problems than it solves. If you limit yourself to EQ'ing only below 80 Hz, a small amount might help a little. Note that professional recording studios tried, and ultimately rejected, EQ about 20 years ago.

--Ethan

 

[Ethan Winer]

Chris,

> a low noise equalizer can be had for $100 <

I have a cat carrier that cost me only $10. This is much cheaper than enough bass traps to properly treat a listening room! But does that make it a good substitute for bass traps? Sure, this is a preposterous reply, but hopefully my point comes through.

> Established perceptual [1]research has concluded that an anomoly of equal magnitude, a null is less objectionable than a peak. <

That research is irrelevant to the subject at hand. Again, the real issue is the key of the music versus the modal response of your room. Here's a graph I love to trot out now and again:

an equalizer does not actually 'add' ringing <

Yes it does! At least it does as soon as you add any boost.

> An equalizer, when dealing with non-delayed resonances, can actually be used to remove ringing. <

Sorry, no. But if you can show an example of this I'd love to see it, and I'll change my opinion on a dime.

> Non delayed resonance will also causes "one note bass". <

Agreed - all low frequency resonances create "one note bass." And this is exactly why EQ is doomed to fail. Unlike bass traps, EQ cannot reduce the damaging low frequency ringing that occurs in all small rooms.

> I would like to read the article, can you please provide the link? <

You bet. Start here, and follow the link to the test results:

www.avsforum.com/avs-vb/showthread.php?t=610122

Of course, the EQ proponent wrote that he believes he defended EQ adequately. But read carefully what I wrote in my "comments and analysis," and look at my examples. If you follow the discussion in the ensuing thread (I hope) you'll come to the inescapable conclusion that EQ might help a little, but it's no match for bass traps.

--Ethan

 

[WmAx]

Chris,

> a low noise equalizer can be had for $100 <

I have a cat carrier that cost me only $10. This is much cheaper than enough bass traps to properly treat a listening room! But does that make it a good substitute for bass traps? Sure, this is a preposterous reply, but hopefully my point comes through.

Yes, it is a prepostorous reply, as it basicly ignored everything I qualifed in the paragraph of mine that it is replying to.

> Established perceptual [1]research has concluded that an anomoly of equal magnitude, a null is less objectionable than a peak. <

That research is irrelevant to the subject at hand. Again, the real issue is the key of the music versus the modal response of your room. Here's a graph I love to trot out now and again:

Your graph is irrelevant. It does not support that assertion that you seem to be making, that a null is more detrimental than a peak of the same magnitude. This is an assertion(unless I have interpreted your statement incorrectly) that is not supported by any research of which I am aware.

> an equalizer does not actually 'add' ringing <

Yes it does! At least it does as soon as you add any boost.

No, it does not add any ringing 'itself'. The ringing is caused by frequency response, whether it is originating from the eq or from the speaker itself. Simple frequency response modification can be used to introduce resonances or remove them when they are minimum phase(non delayed). A room response is not minimum phase, unless the 1/2 wavelength frequency of subject is substantially(wavelength in air) larger than the equivalent room dimensions, because even at 1/2 wavelength, you probably have signficant multiple axis considerations due to posititioning in a non two dimensional world. Of course, if the 1/2 wavelengths were significntly larger than any room dimesion, then it would not be peaking, since it is not subject to modal resonance. If you nontice additional ringing in some specific circumstances by cutting a peak in a room, then that would not be entirely suprising, especially at higher bass frequencies(near and over 100Hz). Realize the speaker is basicly linear(compared to the room). By notching the response, you now have a non-linear transfer function from the speaker. In anechoic analysis, it would be obvious that you have caused energy storage/release problems, at least as actually disperesed from the speaker. Becuase it may sum flatter in pure amplitude response in the room does not do anything to analyze the situation in a satifactory manner. The combination of reflected and direct signal from the speaker can potentially be worse off, so far as time domain, because you have actually de-linearized the speaker's response as a compromise to sum flatter for pure amplitude response in the room. With the right circumstances, ringing is worse.

> An equalizer, when dealing with non-delayed resonances, can actually be used to remove ringing. <

Sorry, no. But if you can show an example of this I'd love to see it, and I'll change my opinion on a dime.

Simple parametric equalization, before/after(this is a moderate modification, I can actually remove the ringing completely if I linearize the response, but this was a practical application to provide users with a solution using 3 bands, not an experiment to try to remove all resonances, which would not be practical and would also decrease power handling of the device) :

Do not confuse non-minimum phase circumstances(like in a room, source signal + delayed reflections) with a minimimum phase circumstance.

You bet. Start here, and follow the link to the test results:

www.avsforum.com/avs-vb/showthread.php?t=610122

Of course, the EQ proponent wrote that he believes he defended EQ adequately. But read carefully what I wrote in my "comments and analysis," and look at my examples. If you follow the discussion in the ensuing thread (I hope) you'll come to the inescapable conclusion that EQ might help a little, but it's no match for bass traps.

Thank you, I will read the link and respond.

-Chris

 

[WmAx]

Ethan, I have reviewed your link. I only concentrated on Front/Center graphs, since this is the primary listening position.

The results are exactly as I would have expected. The magnitude of improvement(<150Hz) as seen in your measurements are small for the bass traps, as compared to the equalization system, until you increase the amount of traps to cover a substantial portion of the room. But it appears that you compromised your effectivness by not corner mounting due to some sort of disagreement with the other person. Even then, the effect of the bass traps at frequencies <50Hz are small. The only conclusion I can make from these measurements is the one I previously had made, which was based on measurements/analysis itself: proper parametric equalization is a more effective means of bass control so far as cost effectiveness and practicality. If the budget is not prohibiting, standard bass traps are obviously far better for midbass, but have little effect into low bass, as compared to equalization. But the conclusion is clear:for optimal behaviour, both equalization and treatments are required.

If you follow the discussion in the ensuing thread (I hope) you'll come to the inescapable conclusion that EQ might help a little, but it's no match for bass traps.

If you modify that statement to read "but it's not match for bass traps above 100Hz", then I'll glady agree.


No EQ/No Traps
http://www.realtraps.com/eq-traps/fc_empty.gif

EQ/No Traps
http://www.realtraps.com/eq-traps/fc_eq-only.gif

No EQ/2 Traps
http://www.realtraps.com/eq-traps/fc_2mondos.gif

No EQ/4 Traps
http://www.realtraps.com/eq-traps/fc_4mondos.gif

No EQ/17 Traps
http://www.realtraps.com/eq-traps/fc_17mondos.gif

And the absolute best response(no suprise), drum roll please....

EQ/17 Traps
http://www.realtraps.com/eq-traps/fc_17mondos-with-EQ.gif

-Chris

 

[mtrycrafts]

,
No EQ/No Traps
http://www.realtraps.com/eq-traps/fc_empty.gif

EQ/No Traps
http://www.realtraps.com/eq-traps/fc_eq-only.gif

No EQ/2 Traps
http://www.realtraps.com/eq-traps/fc_2mondos.gif

No EQ/4 Traps
http://www.realtraps.com/eq-traps/fc_4mondos.gif

No EQ/17 Traps
http://www.realtraps.com/eq-traps/fc_17mondos.gif

And the absolute best response(no suprise), drum roll please....

EQ/17 Traps
http://www.realtraps.com/eq-traps/fc_17mondos-with-EQ.gif

-Chris

Interesting links. I like the EQ only for below92Hz and EQ with all the traps, wonder how much 17 traps cost and the space in the room, above 92Hz. A tough band between 92-128.

 

[bpape]

I think the results may have been quite different with appropriately corner mounted absorbers.


IMO the REASON that peaks tend to be more subjectively objectionable (??) is that in real world situations peaks tend to be more broad - hence more noticed where nulls tend to be somewhat more narrow. However, as Ethan stated, if the null falls on an organ key, drum fundamental, string fundamental, etc. the null can all but obliterate your ever hearing it where the peak would simply exaggerate it.

I'd also like to speak to something Buckle said up a ways. EQ is NOT simply a one seat solution. What you set it to may be based on fixing a single seat but it will change all seats - some for the better, some for the worse.

EQ has it's place. Taming of stubborn peaks AFTER appropriate decay time control has been achieved is a fine place and use for EQ. I highly recommend it. BUT, to attempt to fix all of a room's issues strictly via EQ is not realistic - sorry. In THEORY, you can get some ringing reduction with EQ if you EXACTLY match the inverse of the room function and you only worry about 1 single seat. Neither of those 2 things are ever going to happen. If you don't match it exactly or have more than 1 seat, you see a potential INCREASE in ringing.

Woulld you try to deal with reflection points via EQ only? No. WHY? It's the same thing. It's reflected sound arriving and interacting with direct sound that's messing up frequency response, and causing time related issues. Bass peaks and nulls are no different - you just don't get the 'visual imaging' issues in the bottom end.

Again, using both appropriately is the optimal solution. I'm not saying every room needs 17 or 38 bass absorbers. BUT, every room can stand SOME passive broadband bass control (and passive mid/high control too) to bring the overall decay time of the space in line with projected values based on the usage of the room. THEN deal with the last few gremlins (assuming you've already done seating position, sub position, speaker position, etc.) with EQ that, now, will act more uniformly from seat to seat.

 

[Ethan Winer]

Chris,

> Your graph is irrelevant. <

That graph shows a response that is absolutely typical for every small room. As Bryan and I both explained, if a deep null aligns with a frequency that's present in the music, you will not hear that frequency. If you really believe that not hearing the fundamental (or all-important second harmonic) of a bass note is not audibly objectionable, I don't know what else to tell you. Note that this has nothing to do with those perceptual tests of peaks and nulls you cited because they do not take into account the key of the music.

> No, it does not add any ringing 'itself'. <

Yes it does. Go back and look at any of the "with EQ" graphs and you can clearly see that substantial ringing was added by the equalizer around 92 Hz.

> Simple parametric equalization ... can actually remove the ringing <

That's what Terry Montlick thought too until we did this test. What's missing in your graph is being able to see that the ringing is still present, in the same amount, because it's been pushed below the bottom of the graph. That's why in my analysis section of the report I show the same data with an extra 20 dB at the bottom. Then you can clearly see that the ringing is still present, just reduced in level.

> the effect of the bass traps at frequencies <50Hz are small. <

Agreed. And I'm not opposed to using EQ at the very lowest frequencies. I woudn't do that in my own room, but I wouldn't necessarily criticize someone else for doing that in theirs. Especially if for whatever reason they're not able to install enough bass traps. But once you get above 50 or 60 Hz, EQ is likely to cause more harm than good, and it will not widen modal bandwidth or reduce ringing which are at least as important as having a flat response.

--Ethan

 

[Ethan Winer]

M,

> wonder how much 17 traps cost <

You do not need 17 traps to make a very noticeable improvement! Even four traps, with one in each wall-wall corner, will make a meaningful improvement.

--Ethan

 

[WmAx]

If you really believe that not hearing the fundamental (or all-important second harmonic) of a bass note is not audibly objectionable, I don't know what else to tell you.

I do not know where or how you ever interpreted I said this.

Note that this has nothing to do with those perceptual tests of peaks and nulls you cited because they do not take into account the key of the music.

If you want to make claims, then for accuracy's sake, please present them as speculations, since you can not support your claims of null vs. peak audibility with valid research.

> No, it does not add any ringing 'itself'. <

Yes it does. Go back and look at any of the "with EQ" graphs and you can clearly see that substantial ringing was added by the equalizer around 92 Hz.

*It seems as if you are not reading my replies in their entirety, or just ignoring them.

> Simple parametric equalization ... can actually remove the ringing <

That's what Terry Montlick thought too until we did this test. What's missing in your graph is being able to see that the ringing is still present, in the same amount, because it's been pushed below the bottom of the graph.

You lost me, unless * applies, then your responses make sense. Frequency response is the sole cause of non-delayed resonance( miniminum phase system). It can be introduced or removed with frequency response shaping. Please refer to [1] for information on non-delayed resonance and their relevance. How do you think that,for example, SEAS excel magnesium drivers are used in speaker systems? Their inherant resonance makes them worthless in their natural state, but with an inverse notch filter centered at the resonance(s), they are some of the highest quality midranges available. To suggest otherwise is in defiance of the basic physics involved. If your view is true, then obviously the Excel drivers are not usable in hi-fidelity sound reproduction.

That's why in my analysis section of the report I show the same data with an extra 20 dB at the bottom. Then you can clearly see that the ringing is still present, just reduced in level.

My graph shows the dB floor that is relevant according to perceptual research[1][2]. What you have ignored is that the ringing must still be present in some amount, unless the signal is completely flattened(which can be done, but my graphs are not for that purpose, they are practical application results). Do you notice that the reduction in a band by x amount results in a reduction of the resonance by more than x amount? Do you know how to read or interpret a waterfall/resonance and it's perception relation to audibility? Even if the only the level at which the resonance occurs was changed with a 1:1 ratio(which it is not, becuase resonance is a function of frequency response), then the reduction of level relative to the reference signal is directly relevant to the [1][2]audibility, and lowering it reduces audibility.

-Chris

[1] The Modification of Timbre by Resonances: Perception and Measurement
Floyd Toole, Sean Olive, JAES, Vol. 36, No. 3, 1988, March, pages 122-141

[2]Loudspeaker Distortions, Can We Hear Them?
P.A. Fryer, Hi-Fi News Rec. Rev., vol.22, pp. 51-56, 1977

 

[Buckle-meister]

I'd also like to speak to something Buckle said...

Oh oh...

EQ is NOT simply a one seat solution. EQ has it's place...every room can stand SOME passive broadband bass control (and passive mid/high control too) to bring the overall decay time of the space in line with projected values based on the usage of the room. THEN deal with the last few gremlins (assuming you've already done seating position, sub position, speaker position, etc.) with EQ...

This is what I've been trying to say! I've never said that equalisation shouldn't be adopted. I just feel that starting with absorption is the way to go.

Regards

 

[Ethan Winer]

Chris,

> I do not know where or how you ever interpreted I said this. <

You said:

> Perceptual research demonstrates that a peak is of a higher degree of detriment to sound quality as compared to a null of equal magnitude. <

Then I pointed out that nulls can be even more damaging than peaks in some cases, and that the "perceptual research" you cited doesn't address the situation I'm talking about even though what I'm talking about is very common in all small rooms.

> If you want to make claims, then for accuracy's sake, please present them as speculations, since you can not support your claims of null vs. peak audibility with valid research. <

I have done tons of valid research. Every time I help someone whose mixes had sounded bloated and boomy before adding bass traps, that's another data point.

> Frequency response is the sole cause of non-delayed resonance <

So why do you say that an EQ boost to counter a null will not add ringing? I am all for theory and research, but empirical evidence trumps theory every time. It is very clear to me from those graphs that the EQ added ringing. Is this not clear to you? If not, let me know and I'll point you to the part of the graph that shows this.

> How do you think that,for example, SEAS excel magnesium drivers are used in speaker systems? <

I'm not familiar with those drivers so I have no opinion.

> Do you notice that the reduction in a band by x amount results in a reduction of the resonance by more than x amount? <

Not without being able to see more of the decay at low levels. Do you have another set of graphs that shows more of the decay? Again, I already did that for one of my examples, showing a 40 dB span and then a 60 dB span, at which point you can see that the ringing is still present. I'll also mention that I agree that EQ can reduce ringing a little. But not enough to make a big difference audibly.

> Do you know how to read or interpret a waterfall/resonance and it's perception relation to audibility? <

Sure. I do this stuff all the time.

--Ethan

 

[WmAx]

Chris,

> I do not know where or how you ever interpreted I said this. <

You said:

> Perceptual research demonstrates that a peak is of a higher degree of detriment to sound quality as compared to a null of equal magnitude. <

So, it seems to be confirmed: you are not reading what I write(at least not carefully), as this is different from what you claimed that I said in the last reply.

Then I pointed out that nulls can be even more damaging than peaks in some cases, and that the "perceptual research" you cited doesn't address the situation I'm talking about even though what I'm talking about is very common in all small rooms.

Your graph does not prove that null of equal magnitude to a peak of X frequency, Y gain and Z bandwidth is more audible than peak of X frequency, Y gain and Z bandwidth. Your graph is simply showing some nulls and peaks. Where is the correlary perceptual research?

> If you want to make claims, then for accuracy's sake, please present them as speculations, since you can not support your claims of null vs. peak audibility with valid research. <

I have done tons of valid research. Every time I help someone whose mixes had sounded bloated and boomy before adding bass traps, that's another data point.

That is not research. That is anecdotal, and can not be classified as more than speculation. I can point to any number of people who use the same reasoning("I have done this X number of times so I know") in lack of actual citable research(cable sonics, high resolution audio, immeasurable sound properties of exotic amplifiers, green cd markers, etc.).

So why do you say that an EQ boost to counter a null will not add ringing?

If the dip in response is minimum phase, then it can be countered, and this will reduce ringing, since the drop off edge of a null will result in abrupt phase angle change and itself, be classified as an energy storage mechanism.

I

am all for theory and research, but empirical evidence trumps theory every time. It is very clear to me from those graphs that the EQ added ringing. Is this not clear to you? If not, let me know and I'll point you to the part of the graph that shows this.

Please show me where on my graphs the EQ caused ringing. If you mean your graphs, the low bass was significantly cleaned up with equalization nearly across the board, just look at the low frequency(<92Hz) characteristics.

> How do you think that,for example, SEAS excel magnesium drivers are used in speaker systems? <

I'm not familiar with those drivers so I have no opinion.

It's the same case with nearly all rigid medium drivers, whether it's Focal K, Seas Excell, Eton kevlar, etc.. The stiff structure has a very high Q, very high gain resonance(or resonances) related to the fundamental frequency of the cone.

> Do you notice that the reduction in a band by x amount results in a reduction of the resonance by more than x amount? <

Not without being able to see more of the decay at low levels. Do you have another set of graphs that shows more of the decay?

I will make some new measurements for you within approximately 1-2 weeks.

Again, I already did that for one of my examples, showing a 40 dB span and then a 60 dB span, at which point you can see that the ringing is still present. I'll also mention that I agree that EQ can reduce ringing a little. But not enough to make a big difference audibly.

Why are you confusing your (not minimum phase) graphs with my(minimum phase) graphs? I took the time to explain the differences and how ringing can occur in my #24 post.

> Do you know how to read or interpret a waterfall/resonance and it's perception relation to audibility? <

Sure. I do this stuff all the time.

Which perceptual reference research that you use to coorelate the measured data with audibility?

-Chris

 

[Ethan Winer]

Chris,

> Your graph does not prove that null of equal magnitude to a peak of X frequency, Y gain and Z bandwidth is more audible than peak of X frequency, Y gain and Z bandwidth. <

You are correct. My graph merely shows that all rooms have severe nulls, and common sense then dictates that a severe null will be audible if it aligns with a frequency in the music. Do we agree so far?

> Where is the correlary perceptual research? <

I do this stuff every day, all day long. You can claim it's ancedotal, and perhaps it is, but please don't lump what I'm saying with voodoo and snake oil like green magic markers! But that doesn't make it speculation either because I observe this all the time. The most common problem people have when mixing music in a home studio is deep nulls that make them hear less bass than is really on the track, so their mixes suffer from too much bass. The second most common problem is the opposite: they're sitting in a zone having one or more big bass peaks, so their mixes come out too thin sounding. Are we still in agreement?

> (cable sonics, high resolution audio, immeasurable sound properties of exotic amplifiers, green cd markers, etc.). <

It's clear to me that you and I agree a lot more than disagree, generally speaking!

> If the dip in response is minimum phase, then it can be countered <

Bingo - this is the exact problem. There are so many different reflections in all rooms, all coming from different directions and with different delay times, that no one null can be attributed to a single "filter" pole. So by extension, EQ cannot do much to reduce the ringing because the countering needed is so complex. Perhaps some of the more sophisticated DSP systems could do a better job than a simple parametric EQ, given enough CPU horsepower. But then the problem is that the correction becomes highly localized. Moving even an inch or two - equal to turning your head a little - will cancel the improvement, and surely make the response and ringing much worse.

> Please show me where on my graphs the EQ caused ringing. If you mean your graphs, the low bass was significantly cleaned up with equalization <

Yes, I mean my graphs, and "cleaned up" - however you care to define that - still does not discount the large ringing that EQ added around 92 Hz.

> Which perceptual reference research that you use to coorelate the measured data with audibility? <

Well, I walk into a room and it has huge audible nulls that make the bass all but dissapear. I add a bunch of bass traps and, voila, the bass is nice and round and full again. This happens every time without fail.

--Ethan

 

[WmAx]

You are correct. My graph merely shows that all rooms have severe nulls, and common sense then dictates that a severe null will be audible if it aligns with a frequency in the music. Do we agree so far?

Yes.

I do this stuff every day, all day long. You can claim it's ancedotal, and perhaps it is, but please don't lump what I'm saying with voodoo and snake oil like green magic markers! But that doesn't make it speculation either because I observe this all the time.

[A]For the record: I do not disgaree that nulls are a problem. You, until now, have not paid attention(from what I can tell based on your replies) to what I have taken the time to explain in previous posts and it seems as if you are answering with inapplicable answers/statements to my actual questions/posts.

The most common problem people have when mixing music in a home studio is deep nulls that make them hear less bass than is really on the track, so their mixes suffer from too much bass. The second most common problem is the opposite: they're sitting in a zone having one or more big bass peaks, so their mixes come out too thin sounding. Are we still in agreement?

Well, in the close proximity of the walls that a typical home studio would be set up, then I agree that nulls would probably be the most dominant problem(but this is totally different issue from what I have debated as of this point).

It's clear to me that you and I agree a lot more than disagree, generally speaking!

Perhaps.

> If the dip in response is minimum phase, then it can be countered <

[ B ]Bingo - this is the exact problem. There are so many different reflections in all rooms, all coming from different directions and with different delay times, that no one null can be attributed to a single "filter" pole. So by extension, EQ cannot do much to reduce the ringing because the countering needed is so complex.

I took time to explain these things in great detail in my previous posts, even explaining the relevant variables, but until now you have not acknowledged this.

Perhaps some of the more sophisticated DSP systems could do a better job than a simple parametric EQ, given enough CPU horsepower. But then the problem is that the correction becomes highly localized. Moving even an inch or two - equal to turning your head a little - will cancel the improvement, and surely make the response and ringing much worse.

It's called convulsion. It can be accomplished with DSP(you can completely remove echoes and all manners of non-mininum phase errors). However, as you point out, it is only effective for a very small fixed X, Y, Z coordinate. I am not aware of specific perceptual research demonstrating how small the effective window is, but I would guess that it gives a new meaning to 'head in a vise'.

Yes, I mean my graphs, and "cleaned up" - however you care to define that - still does not discount the large ringing that EQ added around 92 Hz.

The ringing is obviosly an effect that I described in great detail in post #24.

> Which perceptual reference research that you use to coorelate the measured data with audibility? <

Well, I walk into a room and it has huge audible nulls that make the bass all but dissapear. I add a bunch of bass traps and, voila, the bass is nice and round and full again. This happens every time without fail.

It would appear that we are back to [A].

Now that you acknowledged [ B ], do you still want me to make the minimum phase measurements? It seems redundant at this point.

-Chris

 

[Ethan Winer]

Chris,

> I agree that nulls would probably be the most dominant problem(but this is totally different issue from what I have debated as of this point). <

Yes, I have read what you said, and I absolutely agree that what I debated is different from what you were saying. But that's my whole point. I see this often. I'll mention that nulls are a big problem when the null frequency aligns with the music, and someone comes back and cites the research you cited stating it's been proven that nulls are not as audibly damaging as peaks. I then point out that it's an apples-to-oranges comparison, and so it goes.

> It's called convulsion. <

I hope you meant convolution!

BTW, my wife used to work at a hospital years ago and administered ECT. Now that was convulsion!

> I am not aware of specific perceptual research demonstrating how small the effective window is, but I would guess that it gives a new meaning to 'head in a vise'. <

Exactly. A good friend of mine writes exactly this sort of DSP code for remote boardroom conferencing, which is where I learned about all the limitations. A lot of what he does is echo canceling, and he tells me his system is so sensitive it is affected by the person speaking's breathing. As they breath in their chest expands, and that tiny movement reduces his canceling from -50 to -20 dB (or something along those lines).

> do you still want me to make the minimum phase measurements? It seems redundant at this point. <

There's no need for you to do anything unless you'd like to. And I've already seen Floyd Toole's report where he showed that he was able to reduce ringing somewhat using EQ. What's not clear in that report is 1) how long it took him to get that result, and 2) what happens when the measuring microphone is moved half an inch, then one inch, then two inches, etc.

--Ethan

 

[mtrycrafts]

M,

> wonder how much 17 traps cost <

You do not need 17 traps to make a very noticeable improvement! Even four traps, with one in each wall-wall corner, will make a meaningful improvement.

--Ethan

In that graph with 4 traps, I can barely see the changes in the low bands and that is where the ear is also less sensitive to changes, by a whole bunch.

 

[WmAx]

Chris,


Yes, I have read what you said, and I absolutely agree that what I debated is different from what you were saying. But that's my whole point. I see this often. I'll mention that nulls are a big problem when the null frequency aligns with the music, and someone comes back and cites the research you cited stating it's been proven that nulls are not as audibly damaging as peaks. I then point out that it's an apples-to-oranges comparison, and so it goes.

If you would address the statement properly instead of going off into a tangent without warning, it would save you alot of trouble. I'll give an example. I will go back to one of the first posts and I will answer in your place.

I orginally stated:

WmAx: Perceptual research demonstrates that a peak is of a higher degree of detriment to sound quality as compared to a null of equal magnitude. <

You stated in response:

Ethan Winer: That really isn't true, at least not in the context of bass traps and listening to music in a real-life room. A deep null is arguably far more damaging than a peak, at least when the null aligns with the fundamental or 2nd harmonic of a bass note. This depends on the room and the key of the music, of course. A bass null is typically 20 or more dB deep - 35 dB deep is common - and I promise you this is very audible when the null and music are at the same frequency.

Here is how I would have responded(to avoid confusion) if I were you:

I shall clarify: in the context of room induced peaks/dips(nulls); dips tend to be significantly greater in magnitude as compared to the peaks. Specifically, we are not discussing peaks and nulls of equal magnitude.

But instead(if you will review your responses) you apparently decided to debate without giving warning that you were not debating what was being discussed by the other party(me).

> It's called convulsion. <

I hope you meant convolution!

Thank you for correcting my typo.

> do you still want me to make the minimum phase measurements? It seems redundant at this point. <

There's no need for you to do anything unless you'd like to. And I've already seen Floyd Toole's report where he showed that he was able to reduce ringing somewhat using EQ. What's not clear in that report is 1) how long it took him to get that result, and 2) what happens when the measuring microphone is moved half an inch, then one inch, then two inches, etc.

--Ethan

I believe you are referring to Toole's LF correction experiments. I am specfically talking about correction of purely minimum phase device such as a transducer. If you correct for a structure borne resonance on a transducer, there is no problem, as this resonance will exist at all angles relative to the driver in nearly identical ratio to to the reference amplitude response. For example: 6" midrange B has a high Q resonance at 5kHz, and this resonance is at a gain of +8dB relative to the 1kHz reference point. As you move off axis, frequencies that exceed 1/2 wavelength relative to the effective radiating surface will begin to become directive(beam) exceeding 1.5kHz, with significant beaming occuring by 4kHz. The resonance will also beam. A correction made at 0 degrees axis will also be nearly ideal for 40 degrees off axis, and so on, since the resonated output follows the same rules for directivity relative to radiation surface area as the non-resonanted output. This applies to direct radiators. Different dynamic conditions may exist with a horn radiator, for example, where the throat resonances start to deviate from minimum phase behaviour above a certain frequency.

-Chris

 

[Buckeyefan 1]

i looked at the dcx2496 how would i connect it to my receiver as it looks like it has xlr and 1/4" phone plugs can i use an 1/4" to rca adapter without any harm and really i might just use it to smooth the subs out a little thanks

That's about all you'll be doing is using it to smooth out your subs bass response.

You cannot eq digital signals via a surround sound receiver without an external amplifier - and assuming your HT receiver has proper pre-outs.

IE, you cannot run your DVR, DVD coax or toslink into a 2496 eq, then run that to your receiver's digital in. Almost all of todays surround sound receivers lack the tape loop necessary to properly eq 5.1 channels. It's possible to eq a tape deck, or 2 channel cd using analog outs -> cd/tape in on the eq -> output to receiver (composite in).

I think Wmax mentioned it earlier that the correct way to properlly eq all channels would be from the receivers pre out to the eq, then the eq to the amp, then out to the speakers.

(with seperate pre-amp and amplifiers of which you can insert processing between the two)

For all the trouble, you could/should have gone with separates.

FWIW, I am a fan of equalization in 2 channel "analog" modes. For digital surround HT, I think it's great for a dedicated sub. For the mains, center, surround, and surround rear - until manufacturers figure it out, there is no ideal way to integrate any type of eq into a surround sound receiver per se.

 

[WmAx]

I have deleted all posts to brian32672 per our mutual agreement.