Cochlear nonlinearities

[jabbr]

With respect to the question of whether audio frequencies >20 kHz might have an effect on audio perception, it has been often assumed that:

 

1) assume cochlea acts linearly

2) assume basilar membrane does not vibrate >20 kHz due to the anatomy of the cochlea

 

The question really comes down to whether the cochlea acts linearly because if not, then just as high frequency electrical components can cause intermodulation distortion (IMD) so might ultrasonic frequencies affect audio perception.

 

[1] This paper discusses: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6366996/#R30 -- in  particular look at figure 6 -- the thick black line shows response at 22 kHz and presumably greater frequencies!

 

[2] This paper discusses the mechanics of basilar membrane vibration in detail: https://pdfs.semanticscholar.org/6f0a/46befffa8710ebb606ed3dec06d7be317be1.pdf?_ga=2.6977266.115791992.1562285968-561089348.1562285968

 

There are, of course many other scientific papers which bring the notion of cochlear nonlinearities into discussion, and provide mechanisms for ultrasonic effects but these two might give folks a start.

 

The folks who assume there is some absolute law of biophysics that allows the cochlea to be modelled linearly might want to read and understand these in detail --- the situation to me is not so perfectly clear.

 

Feel free to post other scientific studies which either support or refute these ideas (there are many many)

 

[jabbr]

[3] https://www.ncbi.nlm.nih.gov/pubmed/15106276

[4] https://www.ncbi.nlm.nih.gov/pubmed/17691656

[5] https://www.ncbi.nlm.nih.gov/pubmed/9035671

[6] https://www.ncbi.nlm.nih.gov/pubmed/15925206

 

[7] and [8] and for @alfe, an indication that your parent's do in fact  affect your own hearing: https://www.ncbi.nlm.nih.gov/pubmed/29716248, https://www.ncbi.nlm.nih.gov/pubmed/30188326

😂

 

(the salient finding of the above studies is that ultrasonics might benefit tinnitus in some cases -- if so, then this is direct evidence that ultrasonics modulate the auditory system)

[jabbr]

4 hours ago, alfe said:

Just have to wait to be 80 years old and ultrasound will have no secret for me🤫

 

I'm not sure that ultrasound will improve the fuzzy connections between your neurons, but regular exercise and the proper power supply no doubt are important

 

 

[jabbr]

46 minutes ago, adamdea said:

I'm not sure that tinnitus studies have much to do with this.

 

If indeed ultrasonics modulate tinnitus in a human, then this indicates two things:

 

1) ultrasonics are sensed in some way

2) ultrasonics modulate the auditory system.

 

Hard to deny.

[jabbr]

1 hour ago, adamdea said:

I don't think that vaguely referring to nonlinearties in the ear is good enough. Where is the evidence that nonlinearities exsit in the ear which may make hypersonic material in music audible.

 

There are 1098 papers searching for "cochlear nonlinearites" ... this isn't a vague, unsupported concept. Read for yourself:: https://www.ncbi.nlm.nih.gov/pmc/?term=cochlear+nonlinearity 

 

Some of the papers are actually written in a more accessible fashion that might be easier to understand.

 

The technical point is that a Fourier space analysis. e.g. sampling theorem depends on the system being linear -- what I mean when I say this is that each frequency can be considered and treated independently. If the cochlea were liinear that would mean that if there is no response to an isolated tone at 26 kHz, that means that 26 kHz information does not affect the system. Intermodulation distortion is but one type of nonlinearity that is known to affect audio electronics but the cochlea is not a transistor or a microphone or a speaker.

 

Again, the entire analysis that supposedly "proves" that sonic information > 20 kHz can have no affect on audio reproduction because the cochlea does not respond to isolated tones > 20 kHz entirely depends on the assumption that the cochlea behaves linearly. Given the relatively vast amount of research that the cochlea is not linear, I throw out this assumption. Again I am not saying that ultrasonics have been demonstrated to affect music reproduction, nor am I saying that ultrasonics have been proven not to affect music reproduction.

 

In any case the evidence that nonlinearities exist in the ear is contained in those 1098 papers (and elsewhere). Once you accept that there are nonlinearities, you need to throw out your linear assumptions. A wavelet analysis, for example might be more appropriate way to model the cochlea. One example: https://mathematical-neuroscience.springeropen.com/articles/10.1186/2190-8567-1-5

[jabbr]

An interesting perspective from a musical perspective:

 

David Trippett, Music and the Transhuman Ear: Ultrasonics, Material Bodies, and the Limits of Sensation, The Musical Quarterly, Volume 100, Issue 2, Summer 2017, Pages 199–261, https://doi.org/10.1093/musqtl/gdy001

https://academic.oup.com/mq/article/100/2/199/4951391

[jabbr]

17 minutes ago, Rt66indierock said:

 

Good but at the listening position does it matter to a significant part of the market?

 

The vast majority of the market could care less and is happy with AAC or MP3 or whatever happens to be on Youtube

 

Quote

To repeat "You will find high resolution a very hard sell." I wish you luck but I'm pretty busy talking true believers off ledges.

 

I am not selling anything, rather having a hopefully rational discussion.

 

My goal ultimately is to have a more realistic hallucination of a live acoustic performance. I'd also like to improve some live amplified performances. When I say "goal" its not my daily business, rather drives my curiosity.

[jabbr]

1 hour ago, Rt66indierock said:

 

Get me to 1% of the market for anything above CD quality because I can't get there.

 

I can't help you personally with your goal. For a non musician my brain is very close to a musician's. I can easily put myself in Lisner Auditorium when I listen to Waiting for Columbus.  

 

That's what "audiophile" is all about, almost by definition. Very true that musicians will listen to performances, even tryouts, on no more than a laptop. In any case I  listen to Spotify all the time on my iPhone when I'm on the go and its quite enjoyable. I also use a FirstWatt J2 amp as a headphone amplifier which has to be waay  << 1% so its fun to push the envelope.

 

I believe there's room to do some really creative things with deconvolutions and wavelet transforms etc, just as movies are becoming more immersive/VR etc so we shall see what the future brings.

[jabbr]

2 hours ago, Fokus said:

Example: it you press an ultrasonic exciter to your skull (have fun), your will hear something. Does this demonstrate ultrasonic audibility (or rather detectability)? Yes, in a way. Does this rely on intermodulation? Yes, most likely. Is this relevant for audio reproduction in the context we are interested in? No, probably not remotely.

 

 

Thats bone conduction. Bone conduction might indeed be relevant for audio reproduction. 

[jabbr]

27 minutes ago, John Dyson said:

Think about this -- marginally audible sounds at 20kHz (for the fortunate few) are going to produce more distortion effects than anything else for us 'normals'.

 

Think about this: What you call “distortion” is the actual sound including components >20 kHz. 

 

Similar to a tube guitar amp — the sound of the guitar depends on the “distortion” — imagine Jimi Hendrix otherwise.

 

When you filter away >20 kHz you may say that you are removing distortion, but I say you are changing the natural sound.

[jabbr]

2 hours ago, Fokus said:

 

So you listen with speakers pressed hard to your skull?

 

Oh, those interested might want to research what is actually being heard with bone-injected ultrasonics.

 

You are thinking about this in a linear fashion. When I listen to a chorus I don’t hear each individual voice. Likewise listening to a symphony. 

 

Nor or do I listen to each tone nor harmonic. Sounds naturally modulate eachother. 

[jabbr]

1 hour ago, JanRSmit said:

As dat is i onderstand this stuff, it for one confirms that for audible sound levels up to about 50 decibels our hearing system is quite linear. Above that level of becomes increasingly non linear. In the late 70'ties this was already explained to me by medical hearing specialists.

 

Do you mean the response of the cochlea or hearing system in general? For one: what is known about neuroscience is vastly vastly more since the 1970s eg computationally & mathematically. The system is highly nonlinear. Medical hearing specialists in the late 1970s wouldn’t have understood that.

[jabbr]

1 hour ago, JanRSmit said:

How we as human beings sense vibrations is definitely not limited to our hearing system. And how we interpret the vibrations we sense is not only music of noise. It also triggers emotions and body movements like tapping with your feet, etc. 

 

 

 

 

Yes! More than that— low frequencies <20 Hz can be directly felt. This all becomes part of the music experience. 

[jabbr]

59 minutes ago, esldude said:

Any updates since then?

Tinnitus

http://www.tinnitusjournal.com/articles/tinnitus-relief-using-highfrequency-sound-via-the-hypersound-audio-system-10833.html

[jabbr]

1 hour ago, esldude said:

This is production of sound in the audible band via demodulated ultrasound.  Hardly indicative of ultrasound being audible except when being used to produced below 20 khz sound.  The use of bone conduction ultrasound is not new, and also doesn't indicate normal listening is able to hear ultrasound.  

 

I’m not sure you are grokking the full implications of this. 

 

1) Yes it’s possible that, like but not the same as intermodulation distortion, that the mechanism of ultrasonic affect on the cochlea is via this “demodulation” effect — not exactly demodulation but essentially subharmonics. So an ultrasonic signal, via nonlinearities, results in a specific pattern of cochlear excitation. 

2) This pattern might be unique to the individual and specific individual’s response to specific frequencies. 

3) Thus there is no global transform that would allow someone to “reencode” ultrasonic information in the 20-20kHz range. The mapping from 20-40 kHz to 20-20 kHz — or 20-11kHz is presumably specific to the individual. 

4) Thus ultrasonic information is necessary to contain the full experience, as heard by a person. 

 

5) That said it might be possible to “get close” to reencoding 24/96 into 24/44 by modeling an “average” air / cochlear nonlinearity.

[jabbr]

14 minutes ago, pkane2001 said:

And the Tinnitus study is a bit of a stretch,

That’s just one random study of many — point being that nonlinearities can decompose ultrasonic into sonic — that’s all — there are other studies and other mechanisms and other nonlinearities. 

 

How tight do you want to hold on to fixed beliefs despite research?

 

Dogma frequently breaks down when you look at little details.

[jabbr]

11 minutes ago, pkane2001 said:

You referenced the study to make a point, not me 

 

The very specific point I was referencing is that nonlinearities might introduce Sonics as a result of ultrasonics. That’s all. 

 

I am saying that, based on the science we know, it is possible that 24/96 is audibly different than 16/44.1 at least in some circumstances. 

 

I have said many times that I’m not yet convinced either way. I am leaning toward a small difference in some circumstances. 

 

You should get the idea that I don’t adhere to the dogma that 16/44.1 encompasses all that we might ever hear. 

[jabbr]

14 hours ago, marce said:

You would have to wear bone conduction headphones to get any effect, the power of the high frequencies from your speakers is not enough...

 

Certainly the effect of air nonlinearity is seen at 100 - 100 dB SPL. I am giving some known examples of sound produced from ultrasound based on nonlinearities. I don't claim that my examples are exhaustive and certainly a reproduction system which itself filters out ultrasonics would not be able to reproduce.

 

Again, I am not claiming, nor have ever meant to claim, definitive data regarding the audibility of ultrasonic music information, rather casting doubt on the dogma that its not physically possible  that the human can experience anything beyond that contained in the 16/44.1 band. ***

 

Indeed one would expect to need a recording which contains a significant amount of real ultrasonic information and many don't. Apparently the commonly available bone conduction headphones don't have great sound quality for non hearing impaired folks. Complex topic.

 

*** trust me, definitive data would be worthy of a paper in  a journal e.g. AES, we are just shooting the shit here

[jabbr]

6 hours ago, adamdea said:

The first is that whilst no real world device is perfectly linear, we can still treat them as being linear for certain purposes.

 

Yes all known audio reproduction systems are "good enough" for their purpose.