[opaqueice]

Then one must wonder what you're still doing on Audio Circle -- and arguing futilely against established science.

You wonder what I'm doing on AC because I'd rather buy some music than your self-published book?  Logic isn't your strong suit, is it?

So it's really an academic question for most folks then?

Yes.

[opaqueice]

The point is, I think it’s interesting that an objective physiologic measure shows a difference --every time.  But at the same time nobody I know of can actually hear the difference.  Maybe some of you guys can.  I can’t.

I generated an asymmetric test tone once that made it relatively easy for both me and a friend to hear after a little practice.  One polarity sounds slightly higher-pitched than the other. 

I can send it to you or tell you how to generate it, if you'd like.

[stereocilia]

The point is, I think it’s interesting that an objective physiologic measure shows a difference --every time.  But at the same time nobody I know of can actually hear the difference.  Maybe some of you guys can.  I can’t.

I generated an asymmetric test tone once that made it relatively easy for both me and a friend to hear after a little practice.  One polarity sounds slightly higher-pitched than the other. 

I can send it to you or tell you how to generate it, if you'd like.

Only if it's no hassle; of course I'm curious.  Is it a .wav file or some such thing?

 It occurred to me after I thought about it bit more that I have a peripheral memory of a theoretical explanation for the difference in ABR wave morphology between condensation/rarefaction.  I think with condensation the stereocilia bend in one direction to generate a nerve pulse, and for rarefaction they bend the other way, which is intuitive.  Unfortunately, what I don't remember is why that makes a difference.  I'm too close the edge of being a complete bandwidth waste at this point...so...I'll stop here.

[opaqueice]

Only if it's no hassle; of course I'm curious.  Is it a .wav file or some such thing?

I could make one and email it to you if you PM me your address.  Or you can do it yourself - make a test tone consisting of

Code: [Select]

sin[x]+sin[2x+pi/2]+sin[3x+pi]+sin[4x+3 pi/2]+...  You only really need a few terms, but the more you add the more asymmetric it gets (try plotting it and you'll see what I mean).

It occurred to me after I thought about it bit more that I have a peripheral memory of a theoretical explanation for the difference in ABR wave morphology between condensation/rarefaction.  I think with condensation the stereocilia bend in one direction to generate a nerve pulse, and for rarefaction they bend the other way, which is intuitive.  Unfortunately, what I don't remember is why that makes a difference.  I'm too close the edge of being a complete bandwidth waste at this point...so...I'll stop here.

It's because the voltage generated by an inward bend of the cilia is different than for an outward bend.  I saw a talk once where the guy had some plots of the voltage as a function of cilia bend angle (or something equivalent).  To a first approximation they act like rectifiers - they only respond to pressure, and not to rarefaction (or the opposite - I forget). 

Actually it's interesting - the test tone above sounds higher pitched in one polarity than the other, and I think it's because if you look at the waveform you'll see that the positive part is kind of higher frequency than the negative part.  I think the pitch change is due to that and the fact that the cilia responds asymmetrically.

[JerryM]

Only if it's no hassle; of course I'm curious.  Is it a .wav file or some such thing?

I could make one and email it to you if you PM me your address.  Or you can do it yourself - make a test tone consisting of

Code: [Select]

sin[x]+sin[2x+pi/2]+sin[3x+pi]+sin[4x+3 pi/2]+...  You only really need a few terms, but the more you add the more asymmetric it gets (try plotting it and you'll see what I mean).

It occurred to me after I thought about it bit more that I have a peripheral memory of a theoretical explanation for the difference in ABR wave morphology between condensation/rarefaction.  I think with condensation the stereocilia bend in one direction to generate a nerve pulse, and for rarefaction they bend the other way, which is intuitive.  Unfortunately, what I don't remember is why that makes a difference.  I'm too close the edge of being a complete bandwidth waste at this point...so...I'll stop here.

It's because the voltage generated by an inward bend of the cilia is different than for an outward bend.  I saw a talk once where the guy had some plots of the voltage as a function of cilia bend angle (or something equivalent).  To a first approximation they act like rectifiers - they only respond to pressure, and not to rarefaction (or the opposite - I forget). 

Actually it's interesting - the test tone above sounds higher pitched in one polarity than the other, and I think it's because if you look at the waveform you'll see that the positive part is kind of higher frequency than the negative part.  I think the pitch change is due to that and the fact that the cilia responds asymmetrically.

But what does the music sound like? 
It's all about the music.
Have fun,
Jerry

[stereocilia]

I could make one and email it to you if you PM me your address.  Or you can do it yourself - make a test tone consisting of

Code: [Select]

sin[x]+sin[2x+pi/2]+sin[3x+pi]+sin[4x+3 pi/2]+...  You only really need a few terms, but the more you add the more asymmetric it gets (try plotting it and you'll see what I mean).

It occurred to me after I thought about it bit more that I have a peripheral memory of a theoretical explanation for the difference in ABR wave morphology between condensation/rarefaction.  I think with condensation the stereocilia bend in one direction to generate a nerve pulse, and for rarefaction they bend the other way, which is intuitive.  Unfortunately, what I don't remember is why that makes a difference.  I'm too close the edge of being a complete bandwidth waste at this point...so...I'll stop here.

It's because the voltage generated by an inward bend of the cilia is different than for an outward bend.  I saw a talk once where the guy had some plots of the voltage as a function of cilia bend angle (or something equivalent).  To a first approximation they act like rectifiers - they only respond to pressure, and not to rarefaction (or the opposite - I forget). 

It's been a few years since I've done any ABR.  (Can you tell?)  Anyway, movement of the hair cell bundle perpendicular to the direction they are stacked, which I guess could be called a lateral bend, neither depolarizes nor hyperpolarizes the hair cell, that much I'm sure of.  Big deal.  But, I always thought of an inward (toward the kinocilium, let's say) bend of the hair cell bundle as making the hair cell more positively charged by the same amount that an outward bend makes it less positive -- maybe I'm thinking about it incorrectly.  Man, ears are complicated.

[clarkjohnsen]

I can tell you from experience it’s easy to see the effect of polarity when measuring the auditory brainstem response.  However, I would not conclude from this that it is important to preserve it.

Nor would I. I reached my own conclusion from listening for it.

[clarkjohnsen]

But at the same time nobody I know of can actually hear the difference.  Maybe some of you guys can.  I can’t.

On the other hand, everyone I know (anyway everyone to whom I've demoed the phenomenon, and that's not just "audiophiles") can hear the difference. One fellow even wrote me a pretend-hate-mail saying that he can't listen anymore without taking polarity into account -- nor can I.

[stereocilia]

But at the same time nobody I know of can actually hear the difference.  Maybe some of you guys can.  I can’t.

On the other hand, everyone I know (anyway everyone to whom I've demoed the phenomenon, and that's not just "audiophiles") can hear the difference. One fellow even wrote me a pretend-hate-mail saying that he can't listen anymore without taking polarity into account -- nor can I.

I was really just limiting my comment to the square waves commonly used in ABR, the point being that two stimuli that sound the same to me actually show a different physiologic result.  So, somewhere along the line the information is lost by the time it reaches the point where sound is perceived.

We've talked about signals for which discriminating polarity is easily done, but I wonder for which kinds of signals (test tones, or otherwise) are humans unable to do so?  OTOH, I suppose it doesn't matter as long as the signal of interest is music, but the way our auditory system handles sound is really interesting, I think.

[opaqueice]

We've talked about signals for which discriminating polarity is easily done, but I wonder for which kinds of signals (test tones, or otherwise) are humans unable to do so?  OTOH, I suppose it doesn't matter as long as the signal of interest is music, but the way our auditory system handles sound is really interesting, I think.

I can help with that a little bit.  First, only signals composed almost entirely of a fundamental and its harmonics have any chance of having a well-defined polarity, because anharmonic components will destroy the asymmetry.  Second, the signal must be quite asymmetric (by "asymmetric" I mean the positive voltage parts are different from the negative voltage parts) - which means the harmonics must have a phase-shift relative to the fundamental and each other.  Probably one could add more, but those are necessary.

When it comes to natural sounds this means that for many instruments the only possible effect of polarity is on the initial wavefront transient.  For example a triangle or cymbal almost certainly produces a symmetric "ringing" waveform apart from the very beginning.  But some instruments (like brass instruments) make pretty asymmetric waveforms even during a note.

[PMAT]

 

[*Scotty*]

I think after 6 pages this thread finally got to the core of the matter. opaqueice said

When it comes to natural sounds this means that for many instruments the only possible effect of polarity is on the initial wavefront transient.

I have to react with a,(WELL DUHH!) here.  If the music is recorded with attention paid to preserving the polarity of the waveform it may sound more realistic. In an ideal world anyone involved in professionally recording music would take a Hippocratic Oath,ie; to first do no harm to the sound of the music produced during a performance. The thread by extension has made a case for reproducing music with loudspeakers that can recreate a reasonable facsimile of a square wave.  What a concept, not destroying the integrity of a musical waveform in the process of recording it and then reproducing it in your home. A simple idea really,maybe someday it's time will come.
Scotty 

[stereocilia]

I can help with that a little bit.  First, only signals composed almost entirely of a fundamental and its harmonics have any chance of having a well-defined polarity, because anharmonic components will destroy the asymmetry.  Second, the signal must be quite asymmetric (by "asymmetric" I mean the positive voltage parts are different from the negative voltage parts) - which means the harmonics must have a phase-shift relative to the fundamental and each other.  Probably one could add more, but those are necessary.

When it comes to natural sounds this means that for many instruments the only possible effect of polarity is on the initial wavefront transient.  For example a triangle or cymbal almost certainly produces a symmetric "ringing" waveform apart from the very beginning.  But some instruments (like brass instruments) make pretty asymmetric waveforms even during a note.

Okay, so waveforms looking like the virtual trumpet example you linked earlier, http://www.ugcs.caltech.edu/~tasha/Images/myTrumpetC261Zoom.jpg , are periodic yet asymmetric, therefore inverting the polarity will make a difference in the perceived sound.  Is that about right?  I guess once I get the idea that any waveform can be reconstructed from adding the right sine waves together it's difficult to understand why adding together the exact inverse of those same sine waves should sound different.  It's possible that I haven't had enough math training to really wrap my head around this.   Well then, back to my beer and some music with electric guitar distortion!

[opaqueice]

Okay, so waveforms looking like the virtual trumpet example you linked earlier, http://www.ugcs.caltech.edu/~tasha/Images/myTrumpetC261Zoom.jpg , are periodic yet asymmetric, therefore inverting the polarity will make a difference in the perceived sound.  Is that about right? 

Yes.

I guess once I get the idea that any waveform can be reconstructed from adding the right sine waves together it's difficult to understand why adding together the exact inverse of those same sine waves should sound different.  It's possible that I haven't had enough math training to really wrap my head around this. 

It's because each of those sine waves has a phase as well as an amplitude.

If the music is recorded with attention paid to preserving the polarity of the waveform it may sound more realistic. In an ideal world anyone involved in professionally recording music would take a Hippocratic Oath,ie; to first do no harm to the sound of the music produced during a performance.

On natural sounds and music inverting the polarity has very little to no effect the sound.  Humans are really really bad at hearing it, just like we're really bad at hearing phase.  Forcing recording engineers to respect it is a terrible idea, because it would prevent them from using multiple mics (or at least restrict greatly how they can mix them), and many many wonderful recordings have been made using multiple mics.