How much attention do you pay to perceptual research?

[avaserfi]

I am curious to know what many DIYers think about the perceptual research that has been put out by various sources over the years. Many of the principals found to be perceptually important seem to be oftened ignored by DIYers while some others are over emphasized. This makes me wonder what people consider important when they design a speaker and why they think it so?

In my soon to be posted current design I am placing a huge emphasis on perceptual research as I have a friend who has done this previously and has been extremely successful in his builds.

[drf]

The term "perceptual research" is a little benign in my understanding, only because everything to do with sound is perceptual. That then lends itself to the conclusion that it is merely accoustical research.

To be honest, the perceptual research that I've read is an over complicated way of saying we have discovered people hear things in a particular way, so we'll design our products with this in mind. Sounds kinda similar to what most DIY'ers are already doing.

Currently there is already an enourmous wealth of research into accoustics on all levels like psycho-accoustics, material accoustics, room accoustics, etc. To assume that our current understanding of "perceptual research" is essential in speaker design is akin to assuming you can't build a decent amp using bi-polar topology. It is not exactely new or different, simply an alternate way of implementing what we already klnow.

[avaserfi]

Lets take The Modification of Timbre by Resonances: Perception and Measurements. J. Audio Eng. Soc. 36, 122-142, 1988 by Toole, Floyd E. and Olive, Sean E.

There are certain situations which it seems many seem to misunderstand from a perceptual standpoint as they are counterintuitive or just unexpected. For example many believe cabinet resonance is masked by a room, but Toole's perceptual research found that rooms actually increase audibility or resonance. Another is that smaller but wider affected resonances will be less audible when the exact opposite is the truth.

This is a great example of perceptual research as it did not matter what the waves were acting like as measured (except as a quantification methods), but the fact that thresholds for hearing the coloration of sound changes depending on situation. Acoustic research deals with the waves directly perceptual research deals with how we hear the waves as human ears act differently than certain tools.

[drf]

There are certain situations which it seems many seem to misunderstand from a perceptual standpoint as they are counterintuitive or just unexpected.

Who?

For example many believe cabinet resonance is masked by a room, but Toole's perceptual research found that rooms actually increase audibility or resonance.

we already know this, why do you think we spend so much time and money treating our rooms and use only certain timbers and techniques when building speakers?

what exactely is it about perceptual research that you think is different from our current understanding?

[F1 fan]

drf wrote:

we already know this, why do you think we spend so much time and money treating our rooms and use only certain timbers and techniques when building speakers?

Exactly.

[avaserfi]

drf wrote:

we already know this, why do you think we spend so much time and money treating our rooms and use only certain timbers and techniques when building speakers?

what exactely is it about perceptual research that you think is different from our current understanding?

I understand what you are saying I am just asking questions, is that such a bad thing?

If perceptual research was never done how would we know what frequencies ranges humans are most susceptible to timberal distortion on so we could take the appropriate action? Same goes for the fact that a perfectly linear FR is not perceived as linear hence the use of the Toole curve or the type of curve Linkwitz implements so help for designing crossovers was found here. There are countless other examples of what perceptual research has given us in regards to speaker design.

Its the small things many take for granted, but doesn't understanding why and how these things are determined allow for better speaker design? For example the average DIY making a closed cabinet design does not use enough bracing/dampening to reduce timberal distortion to inaudible levels especially when account for the room interaction. Or the fact that even with all this understanding you claim everyone has why few can answer simple questions like what is the perceived effect of timberal distortion? Generally this problem is perceived as warmth. Simple question simple answer but few understand how this happens - how to appropriately deal with it - and why it occurs.

I am not trying to attack anyone, but I am just asking simple questions. I see no need for defensiveness. A proper understanding of how we hear and why we hearing things the way we do seems like it should be imperative to designing loudspeakers that produce sound naturally to our ears - not to a computers label of perfection.

Edit: I should also point out that I am not saying ignore "harder" scientific research for perceptual research, but am saying both should be looked at and accounted for so that the perceptually best speaker possible can be made.

[drf]

It's not a bad thing to ask questions, and I'm sorry if my response came across as defensive (it wasn't) however if you ask a question you should be ready for the answer.

I am confused about why you think everyone is underbracing their designs. Perceptual research, as far as I have read, does not quantify the bracing requirments for specified enclosure design.

A quote from toole himself:

"A properly interpreted set of loudspeaker measurements correlates very strongly with subjective listening impressions."

The diy enthuisasts (just like the professionals) rely heavily on both what sounds good and what measures good. I fail to see the difference between what most of us are currently doing and what toole was researching many years ago. Perceptual research may be the foundation of todays speakerbuilding philophies or it may just be the inevitable science behind what we are doing. But I can garuntee its not new and its not something we are all ignoring. Afterall the basis of tooles research was to correlate loudspeaker measurements with listening impressions, isn't that what we do with every design and prototype?

[avaserfi]

drf wrote:

It's not a bad thing to ask questions, and I'm sorry if my response came across as defensive (it wasn't) however if you ask a question you should be ready for the answer.

I am confused about why you think everyone is underbracing their designs. Perceptual research, as far as I have read, does not quantify the bracing requirments for specified enclosure design.

A quote from toole himself:

"A properly interpreted set of loudspeaker measurements correlates very strongly with subjective listening impressions."

The diy enthuisasts (just like the professionals) rely heavily on both what sounds good and what measures good. I fail to see the difference between what most of us are currently doing and what toole was researching many years ago. Perceptual research may be the foundation of todays speakerbuilding philophies or it may just be the inevitable science behind what we are doing. But I can garuntee its not new and its not something we are all ignoring. Afterall the basis of tooles research was to correlate loudspeaker measurements with listening impressions, isn't that what we do with every design and prototype?

You bring up some very good points about Toole's research, and I do realize that there was never a quantification of the proper amount of bracing. One thing to note about Toole's research is how he controlled for various confounding/extraneous variables which is not done with the majority of normal listening tests. Thusly, the two cannot entirely be compared - Toole mentions this in his research articles. In much of the done research a control is used so that listeners can ABx specific variables, since we are talking about resonance I will continue with that example. In the previously mentioned research article listeners would compare a signal that had resonance to one that didn't in the effort to determine thresholds for hearing such coloration.

Continuing with this example: Very few people do proper measurements to test for cabinet resonance and few own a non-resonant control to compare their builds to [remember resonance does not show up in the normal frequency response graph - I can explain further if needed]. The reason I say many to most builds are under braced is because of my experience involving cabinet resonance. I have discovered it requires extreme bracing/dampening to remove all audible signals which few are able to do because in all honesty it is a PIA to deal with and many do not realize the extent measures need be taken for full removal of this problem.

I think you have also pointed out one semantic error in my OP perhaps ignoring is not the best choice of words, although I am not sure what is best. Also, I am ready for the response and enjoy this discussion greatly . So thanks for entertaining me with it!

Edit: One thing I should note is I am referring to 100% transparent speaker design which perceptually does not alter the source in the slightest unless a PEQ is added to the line and the coloration is added as part of preference.

[drf]

Do you have any links or figures concerning the perceptability of resonance?

It would be beneficial to know exactely what the average level of resonance coloration is present before it becomes percievable. One stumbling block I forsee with this is that in order to percieve whether your enclosure is resonant free or not, would require that you know what the drivers sound like (or measure like) without cabinet or room interaction, which makes prototype and cabinet design difficult, and could require the use of an anechoic chamber for the intial design stages.

[remember resonance does not show up in the normal frequency response graph - I can explain further if needed]

I understand this, this is why I study a wav graph which shows any change in signal not just level and freq'. In this thread (ignore that it is for an amp) you can see the signal that is feed into the system and the resultant wav measured from the speakers. This gives me a good idea of how the system as a whole (including room interaction) is effecting the sound.

[avaserfi]

drf wrote:

Do you have any links or figures concerning the perceptability of resonance?

I do not have a link but I can recommend an article to read. The Modification of Timbre by Resonances By Floyd Toole. JAES - Volume 36 Number 3 pp. 122-142; March 1988.

drf wrote:

It would be beneficial to know exactely what the average level of resonance coloration is present before it becomes percievable. One stumbling block I forsee with this is that in order to percieve whether your enclosure is resonant free or not, would require that you know what the drivers sound like (or measure like) without cabinet or room interaction, which makes prototype and cabinet design difficult, and could require the use of an anechoic chamber for the intial design stages.

The easiest method for testing for cabinet resonance is use of an accelerometer on various points on the outside of the cab then roughly consider the total radiation of the surface area. While this method is not 100% precise it is a good enough estimate to tell.

In all honesty when I evaluate a loudspeaker the last thing I concern myself with is my ears until the actual measurements have been taken. While I realize this is very different from most - with the appropriate graphs one can tell how a loudspeaker will sound without ever having heard it assuming a properly treated environment (Yes I am 100 percent serious).

drf wrote:

I understand this, this is why I study a wav graph which shows any change in signal not just level and freq'. In this thread (ignore that it is for an amp) you can see the signal that is feed into the system and the resultant wav measured from the speakers. This gives me a good idea of how the system as a whole (including room interaction) is effecting the sound.

Actually, I was referring to the fact that typical measurements are taken using gated impulse methods (in attempt to 'remove' the room from the situation) while to see cabinet resonance on a graph one would need to use an impulse response method - far field, at various angles in an anechoic chamber. Once this is done a waterfall can be derived that will show the coloration.

[drf]

So everyone else knows what I am talking about Avaserfi sent me a document study by Floyd Toole and Sean Olive concerning resonance and its hearing threshold. I am only half way through it and will read it again later to make sure I understand it before commenting too much on it. What I will say for now is that I believe it is a lot effort to go through for what could be perceptably insignificant gains, this of course meaning that to the end user it might sound better with a little cabinet resonance.

To put this into perspective, the way I approach any form of audio is to try and match a live quartet playing in chamber. If you ever get the chance to listen to a live quartet (or quintet is better) you will get an idea of what to aim for when setting up an audio system. Using this as your benchmark and letting your ears do most of the prototyping work things like resonance, harmonic distortion, delay and combing as well as room interaction get factored into the design stage and come out the other end as a non issue or best case scenario.

more to come when I have finished said article.

[avaserfi]

drf wrote:

So everyone else knows what I am talking about Avaserfi sent me a document study by Floyd Toole and Sean Olive concerning resonance and its hearing threshold. I am only half way through it and will read it again later to make sure I understand it before commenting too much on it. What I will say for now is that I believe it is a lot effort to go through for what could be perceptably insignificant gains, this of course meaning that to the end user it might sound better with a little cabinet resonance.

I am curious to know what part of the article leads you to believe this? While I am more sensitive than most to resonance I have found that even subtle differences in resonance are detectable by the few people I have been able to drag to the local audio shop and compare resonant speaker to a pair of B&W801D/802Ds which are completely inert - in some cases side by side comparisons were made.

drf wrote:

To put this into perspective, the way I approach any form of audio is to try and match a live quartet playing in chamber. If you ever get the chance to listen to a live quartet (or quintet is better) you will get an idea of what to aim for when setting up an audio system. Using this as your benchmark and letting your ears do most of the prototyping work things like resonance, harmonic distortion, delay and combing as well as room interaction get factored into the design stage and come out the other end as a non issue or best case scenario.

more to come when I have finished said article.

While it is a great goal to recreate real life from memory. The human memory for sound lasts about three seconds, thus it is fairly important to have a reference around when testing the outcome of your builds. The problem is nearly all references available that are inert cost thousands of dollars. Also, another issue here is few have heard inert cabinets and don't always realize it when they do, thus few realize what they are hearing when they hear it.

I look forward to your response and your final thoughts on the paper. In addition to this one, if you have time I strongly suggest reading Loudspeaker Measurements and Their Relationship to Listener Preference Part 1 and Part 2 which I unfortunately don't have electronic copies of. It seems much of Toole's perceptual research builds off of itself in some ways such that one paper doesn't give you the whole story, but all do.

[avaserfi]

thylantyr wrote:

If you ask me, the enclosure design & build would be the easy part as long as you use some common sense and have decent design and woodworking skills. The harder part is finding the cool drivers for the project and trying to integrate them to do the job well [ie, driver locations, crossovers, etc], then you have room interactions that some folks place high on their priority list.

The driver is an extremely important part of the loudspeaker system, but the cab is as well. A weakness in either limits the overall quality. Also, without having a properly treated room having the best speaker in the world is pointless.

All I am pointing out with the cabinet resonance example is that bracing is often not sufficient to remove audibility of resonance in the majority of speaker designs.

thylantyr wrote:

I know that some people also cripple their sound system on the electronics
side of things, ie not using proper amplification for the design, etc.

Very true.

thylantyr wrote:

One thing for sure and I would bet money on this. Those high end 'fancy' brands
using exotic materials for cabinet design 'special glass', 'secret XYZ materials',
will charge alot of money for their products. To justify the high cost, they need
to show that their speaker have something unique and this may be there more
for marketing reasons than sound.

I agree, there is no need for exotic material such as special glass in cabinet design - there is a large amount of marketing B.S. in the world of audio. One should realize that just because things have been done one way for a long time doesn't make them right, there might be a better option that is slightly more complex. This option might not be viable for everyone, but for the few willing to do so - why shouldn't they?

thylantyr wrote:

If you ask any competent DIY speaker designer whos been in this game for along
time, the reason they DIY is to get a better speaker for the money. In other words,
they will build a competent speaker using common sense methodologies with ordinary materials and the end product is better than a store bought product at any price range because it takes more than a box to ruin a design.

Why do you think I am a DIYer? I can't afford the few commercial speakers I would actually want to own. Also, it take more than perfect drivers to have a great speaker.

Thanks for your input thylantyr, hopefully it goes better than our last communication.

[drf]

My sonic memory is about 7 seconds. What I was trying to say is that I have a benchmark for quality that I try to achieve, whether my brain is good enough to compare 2 sounds that are months apart or not doesn't really detract from my ears ability to hear clarity.

More to come gotta go to work.

[avaserfi]

drf wrote:

My sonic memory is about 7 seconds. What I was trying to say is that I have a benchmark for quality that I try to achieve, whether my brain is good enough to compare 2 sounds that are months apart or not doesn't really detract from my ears ability to hear clarity.

More to come gotta go to work.

You are above average and crazy enough to test it, like me . Thank you for the clarification that makes much more sense to me although I still attempt to rely on A/B comparison if using ones ears or certain set actual measurements which I put even more preference on.

I look forward to the rest of your thoughts.

[avaserfi]

thylantyr wrote:

Did you understand my last post ? I linked some manufacturers that want
cabinet resonance in their designs, to give that design it's sonic character.

Yeah, I got it. I am referring to creating a speaker that is completely transparent and nothing else.

thylantyr wrote:

Isn't it then safe to say that removing cabinet resonance isn't
a universal goal for all speaker designs? The evidence seems to
imply that it's a designer preference, not a mandatory requirement.

This I can agree with - if is very safe to say, but in the creation of a completely transparent speaker it is not possible to allow for audible resonance in the design.

thylantyr wrote:

Think musical instruments and their design and wood choice adds to
their sound. Some manufacturers want to add some of this spice
to their design.

I understand that and to be honest I have absolutely zero interest in those speakers as I have previously said. One part of my view of a properly designed is that is imparts absolutely nothing on the source if this is not ideal for an individual my recommendation is an EQ along side this transparent speaker. This is simply because Equalization can be catered to anyone while with cabinet resonance the coloration is set in the speaker.

I realize my goals aren't everyone else's and I have no problem with that, but I would like to point out that having a speaker that imparts absolutely no coloration on the source allows for complete control of sound via a quality equalizer rather something so permanent.

In the end we both have the same goal - to create the best speaker possible - we just go about it in different ways. I am just curious to see how people go about it. I clearly have the stand point that research done in the field of perception is extremely important - due to my psychological background. I also find harder sciences as extremely important as well. Rest assured curiosity causes my questions not me pushing my ways on anyone. Apologies if it comes off that way.