Testing whether the speakers work as intended?

[Matthew J Poes]

I read what Dr. Toole wrote about the same thing too, but I wouldn't take it out of context. I also would bet that the author of REW may disagree with you, as do other experts with PhD degrees in related fields. I know many people are parroting this thing above "don't eq above the Schroeder frequency blablabla...and citing reasons cited by Dr. Toole and others, but if guess if confronted, Dr. Toole will likely qualify his statements. I am not saying it is not an important point to consider, but like a lot of things, it depends..



That sounds fair, but I can claim the same about req software, lots of time people say they don't work, but really? Or just sometimes people don't know how to use them to benefit from them? You seem to be an expert on REW, so you must know if you compare FR plots with Audyssey (just an example, could be Dirac, Anthem's or REW' too) on and off, you may see that with it off, the bass may have some strong bumps that may just please the bass addicts who in turn may claim REQ took the life and punch right out, and that everyone should be better off turning them off, not realizing that there are people who prefer flatter FR including the low end.



You may want to email JA at Stereophile about that too.



Imo that's a valid point, and I would like to watch your video.

Happy holidays!

I’m happy to explain in more technical terms why i feel the way I feel. Know that these ideas aren’t not just blanket opinions of a handful of experts that I took at their word. In fact my first experience with this view came from Earl Geddes who explained his rationale in more specific and technical terms than did Toole. I later took some acoustical physics and acoustical engineering courses while in graduate school (for an unrelated degree) and learned quite a bit more.

JA uses waterfalls in measurements designed to remove the effect of the room. His waterfalls show only the decay behavior of the speaker and not the room. If any of the speaker behaves in a non-minimum phase manner then the waterfall might be useful. That’s actually pretty rare and unlikely so I would still contend their inherent resolution limitations makes them fairly useless. In room waterfalls are completely useless as far as I am concerned. They never show you anything that isn’t already evident in the steady state response. They have inherently poor resolution. That is why I think a wavelet makes more sense if one really wants to use them. They even show you the systems group delay as well as ringing. JA and I had a long talk about this via email after he attended my talk at AXPONA and I didn’t get the impression he disagreed with this. Further, he specifically told me that he tried REW and eq and it wasn’t for him. All Rubinson is more into that it seems. JA not so much.

As for John, you should go through his manuals and posts on eq. He has come right out and cautioned about eq. He thinks it’s greatest value is primarily below Schroeder with careful eq ok in the transition zone. There is no hard cut-off on this but he and I have consistently given people similar guidance. Somewhere between 300 and 500hz is the limit. By this point you are well into the stochastic zone. If you look into what is in a measurement in the stochastic zone it should be fairly obvious why you wouldn’t want to eq that. Its not just that it’s stochastic in nature, it’s also a problem with the measurement technique.

As for PhD’s disagreeing with this view. That is possible but I know of none who specifically study this topic and have a different opinion. I’ve heard variations on this idea but all hold a similar view as to the acceptable way to eq above the 300-500hz threshold. Some feel no eq above Schroeder at all, some feel a modest amount up to 300-500hz could be ok. Some feel averaging is best, some feel minimum deviance is best. That is certainly debated.

 

[JerryLove]

I want to make sure I understand; because I continue to have difficulty getting these statements into a consistent position.

Do rooms alter the sound of speakers (emphasizing or de-emphasizing frequencies)? I'm assuming the answer is "yes".

Should you modify the signal to counter these influences? The statement I keep seeing is "no".
Should you omtimize the position of your speakers to reduce these influences?
Should you add room treatments to reduce these influences?

Unless you answered "no" to all three: what's the functional difference between re-positioning a speaker to get rid of an unwanted sound and modifying the output to get rid of an unwanted sound.

I mean: I agree that removing the cause is likely more effective than adjusting for the cause (turning off a noise is superior to noise cancellation after all); but the idea seems fundamentally the same.

And I remain convinced that not hearing the room is "better"; but perhaps that's less "I'm right" and more "my preferences vary from the norm". Still. I don't believe I'm alone. I think "hearing the room" is a superficial way to 1) correct for improper recordings/playback and 2) like loudness/dynamic compression, sounds good at first but not so much in the long-run.

Then again: I have a history of listening to headphones while sitting next a subwoofer.

 

[Verdinut]

I want to make sure I understand; because I continue to have difficulty getting these statements into a consistent position.
Then again: I have a history of listening to headphones while sitting next a subwoofer.

The vibrating bass which is missing with headphones is then compensated by that of the subwoofer. That must represent a good overall experience.

 

[shadyJ]

I want to make sure I understand; because I continue to have difficulty getting these statements into a consistent position.

Do rooms alter the sound of speakers (emphasizing or de-emphasizing frequencies)? I'm assuming the answer is "yes".

Should you modify the signal to counter these influences? The statement I keep seeing is "no".
Should you omtimize the position of your speakers to reduce these influences?
Should you add room treatments to reduce these influences?

Unless you answered "no" to all three: what's the functional difference between re-positioning a speaker to get rid of an unwanted sound and modifying the output to get rid of an unwanted sound.

I mean: I agree that removing the cause is likely more effective than adjusting for the cause (turning off a noise is superior to noise cancellation after all); but the idea seems fundamentally the same.

And I remain convinced that not hearing the room is "better"; but perhaps that's less "I'm right" and more "my preferences vary from the norm". Still. I don't believe I'm alone. I think "hearing the room" is a superficial way to 1) correct for improper recordings/playback and 2) like loudness/dynamic compression, sounds good at first but not so much in the long-run.

Then again: I have a history of listening to headphones while sitting next a subwoofer.

One thing to remember is that many speakers are designed to be used in room and are engineered to mitigate and also leverage room acoustics effects. They were also designed to be used for certain positioning. For an example, you would not want to place your standard tower or bookshelf speaker right up in a corner or against the wall as that will produce a very audible cupped effect in the sound. But some speakers are engineered to compensate for that, such as on-wall speakers. Also, many tower speakers don't have a flat bass response down to their tuning frequency, they can have a gentle slope starting at like 80 or 90 Hz or so because the designers know the room will boost the low end to some degree. It's the same reason why a uniform off-axis response is advocated so much, so that the acoustic reflections correlate to the direct axis response.

Something else to keep in mind is that not all room acoustics is bad; it's not all unwanted sound. As Floyd Toole says, some reflected sounds are good. They help to enhance the spaciousness and present a broader soundstage. Floyd Toole touches on this in this article. Understand that the music mixes were not made in a totally dead sound console room either. And even many headphones have a frequency response that mimics a room slope; people generally do not like a flat frequency response devoid of room acoustics. I would bet that you wouldn't like to hear the sound of a flat response device that has nothing to temper it. You ought to read the latest edition of Toole's book, he does talk about all of this and cites much of the latest research.

 

[Matthew J Poes]

I want to make sure I understand; because I continue to have difficulty getting these statements into a consistent position.

Do rooms alter the sound of speakers (emphasizing or de-emphasizing frequencies)? I'm assuming the answer is "yes".

Should you modify the signal to counter these influences? The statement I keep seeing is "no".
Should you omtimize the position of your speakers to reduce these influences?
Should you add room treatments to reduce these influences?

Unless you answered "no" to all three: what's the functional difference between re-positioning a speaker to get rid of an unwanted sound and modifying the output to get rid of an unwanted sound.

I mean: I agree that removing the cause is likely more effective than adjusting for the cause (turning off a noise is superior to noise cancellation after all); but the idea seems fundamentally the same.

And I remain convinced that not hearing the room is "better"; but perhaps that's less "I'm right" and more "my preferences vary from the norm". Still. I don't believe I'm alone. I think "hearing the room" is a superficial way to 1) correct for improper recordings/playback and 2) like loudness/dynamic compression, sounds good at first but not so much in the long-run.

Then again: I have a history of listening to headphones while sitting next a subwoofer.

There is a lot here. I’ll touch on just Eq for now. I’m not saying that preconditioning the signal is all bad at all frequencies.

If you go back to the room and the way that modes operate at different frequencies in the room, then in the “modal” resonant zone, I think eq is not only a good idea but necessary.

Between that point and up to the point of clear stochastic behavior, what I call the transition zone, I think eq is ok if used carefully.

Above that point in the stochastic zone I, Geddes, Toole, John (REW), etc. all feel eq is a bad idea. We feel that way for the same reason. My upcoming Audioholics article will do a better job explaining aspects of this. Briefly, what your ears and brain hear isn’t the same as what the mic picks up in the stochastic zone. At those high frequencies the directionality in the ears vs an Omni mic matters. The mic can’t discern reflections from direct sound nor from other reflections other than based on their delay. Our ears can. If a speaker had a flat anechoic response and an ideal polar response (and there was no source of diffraction, reflections, or resonances near the speaker) then eq would be fine. If not (and typically it is not true that a speaker has a flat axial response, perfect polar response, and none of these early problems around the speaker) then the measurements muddle all of this together. Our ears will separate some of them out, filtering out some of the effects. If you EQ these errors then you have applied this potential response error across the board. You may be fixing a problem the mic picked up but isn’t really there, and making things worse. You can’t conflate what the mic picks up with what we hear. The nature of a steady state in room response is not as simple as it seems. It is a complex sum of the speakers direct response, early reflections, diffraction, and resonances all combined.

Now you might ask what is so wrong with Eqing reflections and diffraction. The reason for this would be that if you move the mic even an inch the nature of that error changes. A hallmark of a diffraction distortion is that the peak or null moves in frequency as the angle changes. If you took 100 measurements around a small area the size of your head, you would likely find that the diffraction effects mostly average out. The best fix for a diffraction effect is to get rid of it at the source. It can’t be eqed.

Ok what about reflections at high frequencies. Well, these cause modal fluctuations much like at low frequencies. The difference is that they move around over short distances. These modes lie right on top of each other from all the reflections. If you eq based on a single point in space, the eq would be wrong. Our ears are not in a single point in space. If we take an average it still is equally accounting for reflections in all directions. Our ears are forward biased and filters out early reflections. That would make eq problematic too.

HF resonances are something like a glass door near the speaker hitting a coincident frequency. Like diffraction, it’s better to just damp the resonance rather than trying to eq it out. These are also fairly uncommon.

It isn’t that it’s impossible to take in room measurements and eq at HF’s accurately. It’s that you can’t do it the way most people do. REW can’t do it. It requires significant analysis of a series of spatial measurements to obtain appropriate eq. The eq also needs to be fairly special. DIRAC is one of the systems I think does a good job with this. Geddes and Toole don’t seem confvicned, but I have had good experiences with it and the science behind the method has made good sense to me. It’s also clear that REW and PEQ can’t replicate what DIRAC is doing.

https://pdfs.semanticscholar.org/faac/9d619123c532938a514c77f966df10f03e91.pdf

https://www.researchgate.net/publication/224374533_Spatially_Robust_Audio_Compensation_Based_on_SIMO_Feedforward_Control

http://guillaume.perrin74.free.fr/ChalmersMT2012/Papers/Inverse Filtering/IEEE_57_2009_SIMO_AudioCompensation.pdf

And Dirac is developing a MIMO based correction system that I think holds huge promise to revolutionize how in room setup and correction can be optimized and automated. You talk a lot about not wanting to hear the room. That’s a great thing, remove the room to hear the recording as intended. It’s not really possible thought. If a 2 channel system is played back in a room with no reflections it would not sound good (yes Toole tested that, but lots of other work has shown people freak out in anechoic rooms). To recreate all of the acoustics of the original recording requires a multichannel system and to fully mitigate the natural acoustics of the room is really difficult to do. A MIMO system actually has the capacity to change a rooms acoustics to whatever is necessary, including eliminating its natural acoustics through signal processing and lots of speakers. It’s pretty cool. It’s best described as active room acoustics.

 

[JerryLove]

For an example, you would not want to place your standard tower or bookshelf speaker right up in a corner or against the wall as that will produce a very audible cupped effect in the sound. But some speakers are engineered to compensate for that, such as on-wall speakers.

A good example of something that can be compensated for with exactly the sort of EQ that is being argued against in this thread.

Introduce a baffle (wall) and we are adding what, 6db, to frequencies that reflect off that wall? How is fixing that possibly a bad thing?

Something else to keep in mind is that not all room acoustics is bad; it's not all unwanted sound. As Floyd Toole says, some reflected sounds are good. They help to enhance the spaciousness and present a broader soundstage. Floyd Toole touches on this in this article.

From the article:
"Reflections within listening rooms are real and numerous. Some would argue that they all are problems to be eliminated. Others take a more philosophical view that they just provide information about the room, and the brain can figure it out. I’m somewhere in the middle, but leaning towards the latter. "

Again: if reflections are needed then most room acoustic treatments are "bad"; since that's what they cut down upon.

Understand that the music mixes were not made in a totally dead sound console room either.

I am aware. it's part of my theoretic concern with introducing room echos and sonic changes. You are adding a room to a room.

And even many headphones have a frequency response that mimics a room slope; people generally do not like a flat frequency response devoid of room acoustics.

Understand that music mixes were not made in a totally dead sound console room. There are room acoustics baked in to the recording.

There are no headphones that mimic the particular room you happen to be in when listening.

Again: we hit a logical wall.

If an accurate playback of an accurate recording is "bad", then the original sound must also be "bad".

What is the theory in which the master recording "sounds bad" and needs to have "a room slope" added to it, and also more echos, in order to sound good?

I would bet that you wouldn't like to hear the sound of a flat response device that has nothing to temper it. You ought to read the latest edition of Toole's book, he does talk about all of this and cites much of the latest research.

I would take that bet and I'm reasonably sure I'd win it. Though, of course, it depends on the original recording. I've far too strong a fondness for near-field listening to not know my preference there.
(I should note: flat at my ear. I'm not really sure how my preferred headphones measure. Because of the substantial differences in the environment inside a small can; I would not be at all surprised if a non-flat output is needed to get a flat input because of the nature the sound travels in (like how you are going to get bass boost from a room, and would need to lower your bass output to actually reproduce the recording).

Let's try this with a thought experiment.

I'm sitting in a room listening to a live performance. I love it. We replace my head with a dummy head with mics on both sides and make a binaural recording of exactly the thing I was listening to and loving.

You are telling me that, on playback, I need to add new reflections, and sound coloration to that or I won't like it? How so? I liked it when it was live? What could possibly make an accurate reproduction [at my eardrum] "bad" such that I need an inaccurate reproduction to compensate for it?

 

[Matthew J Poes]

A good example of something that can be compensated for with exactly the sort of EQ that is being argued against in this thread.

Introduce a baffle (wall) and we are adding what, 6db, to frequencies that reflect off that wall? How is fixing that possibly a bad thing?

From the article:
"Reflections within listening rooms are real and numerous. Some would argue that they all are problems to be eliminated. Others take a more philosophical view that they just provide information about the room, and the brain can figure it out. I’m somewhere in the middle, but leaning towards the latter. "

Again: if reflections are needed then most room acoustic treatments are "bad"; since that's what they cut down upon.


I am aware. it's part of my theoretic concern with introducing room echos and sonic changes. You are adding a room to a room.


Understand that music mixes were not made in a totally dead sound console room. There are room acoustics baked in to the recording.

There are no headphones that mimic the particular room you happen to be in when listening.

Again: we hit a logical wall.

If an accurate playback of an accurate recording is "bad", then the original sound must also be "bad".

What is the theory in which the master recording "sounds bad" and needs to have "a room slope" added to it, and also more echos, in order to sound good?


I would take that bet and I'm reasonably sure I'd win it. Though, of course, it depends on the original recording. I've far too strong a fondness for near-field listening to not know my preference there.
(I should note: flat at my ear. I'm not really sure how my preferred headphones measure. Because of the substantial differences in the environment inside a small can; I would not be at all surprised if a non-flat output is needed to get a flat input because of the nature the sound travels in (like how you are going to get bass boost from a room, and would need to lower your bass output to actually reproduce the recording).

Let's try this with a thought experiment.

I'm sitting in a room listening to a live performance. I love it. We replace my head with a dummy head with mics on both sides and make a binaural recording of exactly the thing I was listening to and loving.

You are telling me that, on playback, I need to add new reflections, and sound coloration to that or I won't like it? How so? I liked it when it was live? What could possibly make an accurate reproduction [at my eardrum] "bad" such that I need an inaccurate reproduction to compensate for it?

I think maybe what is missing in that thought experiment is the assumption that the reproduction system can reproduce accurately the original acoustics. If it is 2-channel it can’t. If it is multichannel it can. If it is 2-channel then one way to get back some of those natural acoustics is with your room as a stand in. That seems to be the norm and testing of this showed that it was preferred, per Tooles work mentioned earlier. If it is multichannel, I think even Toole has states that making the room very dead and letting the speakers handle the room sound is best. I do think that is a good idea. It does seem that there is general agreement that multichannel rooms should be deader than 2-channel rooms.

So what is our system in this thought experiment?

I’d also separate all this from the eq discussion. I don’t see the relationship between the two anymore. Not Eqing at high frequencies is specific to the problems in measuring and separating out what can be eqed and what can’t. It’s a measurement problem.

 

[JerryLove]

At those high frequencies the directionality in the ears vs an Omni mic matters. The mic can’t discern reflections from direct sound nor from other reflections other than based on their delay. Our ears can.

I must be missing something you are trying to say. A mic doesn't need to discern anything, simply record it.

There's certainly the potential issue of the lack of a baffel on an omnidirectal mic; but that's a recording problem; and I certainly don't see how anything we've discussed in this thread fixes or exacerbates that.

The nature of a steady state in room response is not as simple as it seems. It is a complex sum of the speakers direct response, early reflections, diffraction, and resonances all combined.

Of course.

Speaker designers, including Toole, already do exactly that.

The speaker cabinet is a baffle for some frequencies. That means that there's an increase in SPL within the output arc of the speaker. In order to get flat @1m measurements, the designers reduce output in those frequencies with the crossover.

If I, say, shove the speaker in a wall up to the faceplate: I'm going to have introduced a functionally infinite baffle (a bad one as it's drywall; but that's nuance). I'm having real difficulty understanding a reason I wouldn't want to do the exact same thing that the speaker designer would have done and drop the output on all the frequencies that were omnidirectional and now no longer are (which has made them louder).

Now you might ask what is so wrong with Eqing reflections and diffraction. The reason for this would be that if you move the mic even an inch the nature of that error changes. A hallmark of a diffraction distortion is that the peak or null moves in frequency as the angle changes. If you took 100 measurements around a small area the size of your head, you would likely find that the diffraction effects mostly average out. The best fix for a diffraction effect is to get rid of it at the source. It can’t be eqed.

I generally agree.

I do feel that there's a lot of daylight between this statement and the one that started this conversation. It seems like we've moved from "EQing above 500Hz is bad because you don't hear in that range" to "there are certain causes of distortion that EQing won't fix because the effect is inconsistent throughout the listening space. Correcting for one spot can worsen another only a few inches away"

If you eq based on a single point in space, the eq would be wrong. Our ears are not in a single point in space. If we take an average it still is equally accounting for reflections in all directions. Our ears are forward biased and filters out early reflections. That would make eq problematic too.

Meaning "we'd need to shove the mic against our eardrum and not move our head once calibrated to get it really right".

HF resonances are something like a glass door near the speaker hitting a coincident frequency. Like diffraction, it’s better to just damp the resonance rather than trying to eq it out. These are also fairly uncommon.

Room treatment > EQ. I agree.

I also generally agree with the rest of the post I am quoting; so won't quote it.

 

[Matthew J Poes]

Headphones also don’t have a flat response. Not even close. Most have the same response tilt that was found to be preferred in-room. Actual flat response headphones in the testing Harman did were not well liked nor do many exist in the market. They also have a pretty crappy response in the highs and most don’t extend very deep in bass. In fact the best bass in headphones at this point is coming from digitally processed headphones as the natural response is often pretty rolled off.

 

[JerryLove]

I think maybe what is missing in that thought experiment is the assumption that the reproduction system can reproduce accurately the original acoustics. If it is 2-channel it can’t. If it is multichannel it can.

I disagree quite strongly; both from theory and practical experience with 3D sound on headsets.

If it is 2-channel then one way to get back some of those natural acoustics is with your room as a stand in. That seems to be the norm and testing of this showed that it was preferred, per Tooles work mentioned earlier. If it is multichannel, I think even Toole has states that making the room very dead and letting the speakers handle the room sound is best. I do think that is a good idea. It does seem that there is general agreement that multichannel rooms should be deader than 2-channel rooms.

I agree that, in a dead room, you are going to want multi-channel. Especially far field.

So what is our system in this thought experiment?

A headset playing a binaural recording.

I’d also separate all this from the eq discussion. I don’t see the relationship between the two anymore. Not Eqing at high frequencies is specific to the problems in measuring and separating out what can be eqed and what can’t. It’s a measurement problem.

OK. I'm good with that. "Should reproduction be accurate" as separate from "can you make useful measurements above 500hz that allow you to EQ a room in a manner that improves the subjective experience".

 

[Matthew J Poes]

Ok @JerryLove i think we might be describing two kinds of diffraction problems. Baffle diffraction is different from what I was talking about. That is a steady rise. There is nothing wrong with Eqing that. That is also largely at low frequencies below 500hz. It still is fitting with my approach. Yes if you shove a speaker into a wall you can and should eq it to get a flatter response (though as @shadyJ said, a tilted response in room has been found to be preferred).

What I’m talking about are the kinds of very Hf diffraction distortions that cause little ripples in the response above say 500hz. These can be caused by all sorts of things. Screws near the tweeter, a coffee table between you and the speakers, the speaker grill frame, etc. As I’m sure you know, diffraction is generically the effect of sound waves wrapping around a barrier. Baffle diffraction is the effect of it wrapping around the entire baffle. But I’m talking about when it wraps around smaller less smooth objects or other objects farther from the speaker. These are the diffraction distortions that tend to contribute to all that in room garbage we measure.

 

[JerryLove]

Headphones also don’t have a flat response. Not even close. Most have the same response tilt that was found to be preferred in-room. Actual flat response headphones in the testing Harman did were not well liked nor do many exist in the market. They also have a pretty crappy response in the highs and most don’t extend very deep in bass. In fact the best bass in headphones at this point is coming from digitally processed headphones as the natural response is often pretty rolled off.

As measured at the eardrum while on a head or at the driver in open space?

And yes, of course there are bass problems with cans (probably why I sit next to a sub sometimes when listening). There are bass problems with speakers as well. Cans tend to go to heck at high treble as well (my sony's fluxuate +/- 10db >10Khz; though they are +/-5db 20hz - 9khz... So perhaps not as bad as all that give the range you and I are discussing starts at 500hz. That feels like a subject change.

 

[JerryLove]

Ok @JerryLove What I’m talking about are the kinds of very Hf diffraction distortions that cause little ripples in the response above say 500hz. These can be caused by all sorts of things. Screws near the tweeter, a coffee table between you and the speakers, the speaker grill frame, etc. As I’m sure you know, diffraction is generically the effect of sound waves wrapping around a barrier. Baffle diffraction is the effect of it wrapping around the entire baffle. But I’m talking about when it wraps around smaller less smooth objects or other objects farther from the speaker. These are the diffraction distortions that tend to contribute to all that in room garbage we measure.

Those are things that have very different impacts throughout even a small listening "sweet spot". I agree that attempting to solve those with EQ is a fools errand and they should be resolved by eliminating the source of the problem.

 

[Matthew J Poes]

I’ll also mention that I wasn’t talking about headphones in my 2-channel comment. Even then, high order 3-d sound fields can only be reproduced accurately through binaural headphone systems of head tracking is used. Standard binaural is only capable of reproducing low order sound fields.

An example paper:
https://www.researchgate.net/publication/278404811_Binaural_Reproduction_of_Higher_Order_Ambisonics_A_Real-Time_Implementation_and_Perceptual_Improvements

 

[Matthew J Poes]

Those are things that have very different impacts throughout even a small listening "sweet spot". I agree that attempting to solve those with EQ is a fools errand and they should be resolved by eliminating the source of the problem.

Ok then we really are in agreement. This is the heart of the stuff that is readily picked up and conflated with the direct sound of the speaker that one would be trying to eq If they took a measurement in room of their system (or even many measurements) and asked REW to generate an eq for the speaker. This is what I was saying is bad.

 

[Matthew J Poes]

As measured at the eardrum while on a head or at the driver in open space?

And yes, of course there are bass problems with cans (probably why I sit next to a sub sometimes when listening). There are bass problems with speakers as well. Cans tend to go to heck at high treble as well (my sony's fluxuate +/- 10db >10Khz; though they are +/-5db 20hz - 9khz... So perhaps not as bad as all that give the range you and I are discussing starts at 500hz. That feels like a subject change.

I meant as measured on a dummy head or headphone nearing rig.

 

[shadyJ]

A good example of something that can be compensated for with exactly the sort of EQ that is being argued against in this thread.

Introduce a baffle (wall) and we are adding what, 6db, to frequencies that reflect off that wall? How is fixing that possibly a bad thing?

From the article:
"Reflections within listening rooms are real and numerous. Some would argue that they all are problems to be eliminated. Others take a more philosophical view that they just provide information about the room, and the brain can figure it out. I’m somewhere in the middle, but leaning towards the latter. "

Again: if reflections are needed then most room acoustic treatments are "bad"; since that's what they cut down upon.


I am aware. it's part of my theoretic concern with introducing room echos and sonic changes. You are adding a room to a room.


Understand that music mixes were not made in a totally dead sound console room. There are room acoustics baked in to the recording.

There are no headphones that mimic the particular room you happen to be in when listening.

Again: we hit a logical wall.

If an accurate playback of an accurate recording is "bad", then the original sound must also be "bad".

What is the theory in which the master recording "sounds bad" and needs to have "a room slope" added to it, and also more echos, in order to sound good?


I would take that bet and I'm reasonably sure I'd win it. Though, of course, it depends on the original recording. I've far too strong a fondness for near-field listening to not know my preference there.
(I should note: flat at my ear. I'm not really sure how my preferred headphones measure. Because of the substantial differences in the environment inside a small can; I would not be at all surprised if a non-flat output is needed to get a flat input because of the nature the sound travels in (like how you are going to get bass boost from a room, and would need to lower your bass output to actually reproduce the recording).

Let's try this with a thought experiment.

I'm sitting in a room listening to a live performance. I love it. We replace my head with a dummy head with mics on both sides and make a binaural recording of exactly the thing I was listening to and loving.

You are telling me that, on playback, I need to add new reflections, and sound coloration to that or I won't like it? How so? I liked it when it was live? What could possibly make an accurate reproduction [at my eardrum] "bad" such that I need an inaccurate reproduction to compensate for it?

It may be that a dummy head recording with binaural mics in a typical room would already have the kind of preferred sound that people like, and so replaying that recording in a typical room on a typical system might sound a bit unnatural. But the thing is, most recordings aren't made that way. And again, human hearing acclimates to room acoustics very quickly, so that might not even be a big deal.

Regarding trying to EQ out the effects of placing a speaker too close to a surface, remember that there will be diffraction added to that as well, which can not be compensated for in equalization.

I get the source of your confusion, that how can the original recording be heard as 'bad' if it is heard in its most uncolored state. I can only say that music recordings aren't intended to be listened to in an anechoic chamber. If you think your headphones have a flat response, you are wrong, at least if those headphones are good. Most people would not enjoy the sound of flat measuring speakers in an anechoic setting. But, as has been mentioned, that is less true to surround sound systems.

 

[JerryLove]

It may be that a dummy head recording with binaural mics in a typical room would already have the kind of preferred sound that people like, and so replaying that recording in a typical room on a typical system might sound a bit unnatural. But the thing is, most recordings aren't made that way. And again, human hearing acclimates to room acoustics very quickly, so that might not even be a big deal.

Regarding trying to EQ out the effects of placing a speaker too close to a surface, remember that there will be diffraction added to that as well, which can not be compensated for in equalization.

I get the source of your confusion, that how can the original recording be heard as 'bad' if it is heard in its most uncolored state. I can only say that music recordings aren't intended to be listened to in an anechoic chamber. If you think your headphones have a flat response, you are wrong, at least if those headphones are good. Most people would not enjoy the sound of flat measuring speakers in an anechoic setting. But, as has been mentioned, that is less true to surround sound systems.

Echos and coloration are two different issues. We seem to have flipped between them regularly.

Speakers being "not flat" have little to nothing to do with echos. As I recall what I have read of Toole's work: the most preferred speakers were those that were both the flattest, and had the flattest off-axis performance.

An anechoic chamber removes echos. I don't honestly know what it does with standing waves and other forms of gain. It does not color the output (and if you like multi-channel in anechoic but not stereo, then your concern is reflections, not color).

A room both adds echos and color. Acoustic treatments attempt to lessen both as much as possible. I have not been an anechoic chamber; but I prefer dampened rooms to undampened ones when listening to speakers. Positioning does become more important once echos are reduced. I also like headphones (which completely remove room interactions).

Color I don't see as possible to be "good". The reason being the assumption that the source was the correct volumes already.

Echo I don't see as good either; though I'm willing to cede a little more room to the fact that what is miked and what reaches your ear, especially with speakers in rooms, isn't an exact match. That is not to say I'm ready to agree. I'm not. I just acknowledge that it's a bit more complex.

 

[shadyJ]

Echos and coloration are two different issues. We seem to have flipped between them regularly.

Speakers being "not flat" have little to nothing to do with echos. As I recall what I have read of Toole's work: the most preferred speakers were those that were both the flattest, and had the flattest off-axis performance.

An anechoic chamber removes echos. I don't honestly know what it does with standing waves and other forms of gain. It does not color the output (and if you like multi-channel in anechoic but not stereo, then your concern is reflections, not color).

A room both adds echos and color. Acoustic treatments attempt to lessen both as much as possible. I have not been an anechoic chamber; but I prefer dampened rooms to undampened ones when listening to speakers. Positioning does become more important once echos are reduced. I also like headphones (which completely remove room interactions).

Color I don't see as possible to be "good". The reason being the assumption that the source was the correct volumes already.

Echo I don't see as good either; though I'm willing to cede a little more room to the fact that what is miked and what reaches your ear, especially with speakers in rooms, isn't an exact match. That is not to say I'm ready to agree. I'm not. I just acknowledge that it's a bit more complex.

Your brain knows what a neutral sound is supposed to be like in room. If you don't have that room, you don't have any acoustic interaction. Few people think that listening to a recording in an environment with no acoustic interaction, i.e., and anechoic chamber, sounds good. You can get something like that sound if you find an open field and set your system up there. It sounds very dry. The target frequency response of many headphones nowadays imitates that of a neutral system in room, not flat. The reason for that is because it is what most people think sounds natural. Much of the research in this area is cited by Floyd Toole in chapter 7 of Sound Reproduction, 3rd Edition.

 

[Kvn_Walker]