Skip to main content

Topic: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio? (Read 10068 times) previous topic - next topic

0 Members and 1 Guest are viewing this topic.
  • krabapple
  • [*][*][*][*][*]
Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
I won't give away the answer  ;)

Nice thorough article  by Archimago.

https://archimago.blogspot.com/2017/04/musings-do-we-need-those-20khz.html

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #1
Well, to put it in perspective, if 70+ yr old Dunning-Kruger audiophiles can "hear" improvements with an amp stand, anything is possible, like their 10Khz ears "hearing" 20+ Khz
Loudspeaker manufacturer

  • Porcus
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #2
Well, to put it in perspective

OK, so the perspective you want to promote, is some internet idiot said ...

I think I'll take it out of that perspective and return to the perspective the OP links to: what sense can we make out of scientific research?

Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #3
In the end, it comes down to the same thing: there are those that believe their imagination is real and those that don't.

Have "heard" those Tannoy supertweaters at a friend's house. His sources were vinyl and CD, not even high-sample-rate digital. I didn't bother to ask him where he thought the >20kHz was coming from. Or if he enjoyed the >20kHz noise that might have been coming from his LPs. Hey, it was a social visit, and I did very much enjoy the audible-range sounds coming from the other drivers in his Tannoy speakers :)

And... 64 and can't even hear my favourite kitchen timer any longer, which my phone app tells me is 4k. So am disqualified from comment on what is or isn't there in the top couple of octaves of normally-audible music. And baby-bird song. And stuff.

Always enjoy Archimago's writing. Probably nothing new and this one and, as ever, the people that need to read it won't.
 
The most important audio cables are the ones in the brain

  • KozmoNaut
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #4
And... 64 and can't even hear my favourite kitchen timer any longer, which my phone app tells me is 4k. So am disqualified from comment on what is or isn't there in the top couple of octaves of normally-audible music. And baby-bird song. And stuff.

You probably have some hearing left above 4kHz, it is very common to have a dip around that frequency. I'm 31 and I have it too, only in my left ear.

I seem to remember reading somewhere that hearing damage around ~2kHz tends to harm ~4kHz more for some reason.

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #5
so the perspective you want to "promote"
Nope, just your imagination, like audiophile 20k hearing.
Loudspeaker manufacturer

  • Palladium
  • [*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #6
"Speaker non-linearities result in intermodulation and subharmonic distortions that may be audible <20kHz. This is reason not to record too much ultrasonic content nor try to reproduce it."

But-but-but if we don't intentionally degrade the signal with stupid then how do we get audio transparency?

Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #7
And... 64 and can't even hear my favourite kitchen timer any longer, which my phone app tells me is 4k. So am disqualified from comment on what is or isn't there in the top couple of octaves of normally-audible music. And baby-bird song. And stuff.

You probably have some hearing left above 4kHz, it is very common to have a dip around that frequency. I'm 31 and I have it too, only in my left ear.

I seem to remember reading somewhere that hearing damage around ~2kHz tends to harm ~4kHz more for some reason.

I'll have to listen to some test tones again, but I find it rather depressing.

But I do not play the audiophile numbers game (which would be even more depressing) because I know there is still plenty of music in the frequencies that I can hear. But cymbal sheen and "air," alas, no.
The most important audio cables are the ones in the brain

  • krabapple
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #8
Well, to put it in perspective, if 70+ yr old Dunning-Kruger audiophiles can "hear" improvements with an amp stand, anything is possible, like their 10Khz ears "hearing" 20+ Khz


pseudoscience is even better when it's all in blue

http://www.whatsbestforum.com/showthread.php?18367-Introspection-and-hyperbole-control&p=444653&viewfull=1#post444653

  • Jakob1863
  • [*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #9
First of all, if talking about pseudoscience or scientific reasoning/rigor at all, is archimago´s writing really on par?

There are at least four different topics to discuss:
1.) does music exist that contains some energy above 20kHz?

As Boyk´s already cited paper shows, the answer has to be yes

2.) do microphones exist that are able to work sufficiently in the range above 20kHz?

The answer has to be yes, although certainly not every microphone used for recording is able to

3.) what about intermodulation distortion?

A valid concern, but if the microphones are source of IM distortions then we have to do some acoustical filtering between source and microphone to avoid IMD. Otherwise we have to realize that IMDs are already and inseparable imbedded in the recorded material.

The discussion about IMD reminds to a quite similar discussion roughly 60 years ago, when Olson tried to figure out if listeners would prefer full range sound or a bandwidth limited version of the "real thing". /1/

Of course high frequency content means higher risk of IMD, but that would only mean to use better equipment that does not suffer from higher IMDs due to extended high frequency content. (if human listener really prefer content including high frequency sound)

4.) do human listeners prefer or at least notice if music does contains energy above 20kHz (during the original event or the reproduction)?

After all these years that is still hard to answer. Up to a certain degree that reflects the general unwillingness to replicate studies done earlier.
Oohashi et al. examined the socalled Hypersonic Brain Effect and tried to get some more objective data by inclusion of PET scans and EEGs, but did a psychoacoustic evaluation as well . /2/

As usual there were much (sometimes heated) discussions about their study, but afaik no one tried to replicate it.
There were several follow up studies, the most recent one published in 2017, done by various experimenters and in every case some confirmation for those hypersonic effects was found. (if somebody is interested i´ll provide some citation for those too)

I do remember that pioneer introduced their socalled music-link digital filter in CD players during the 90s because listeners would prefer the reproduction if some content above 20kHz is included.  (afair according to pioneer based on listening tests)

/1/ Harry F. Olson, Frequency Preference for Speech and Music,  The Journal of the Acoustical Society of America 19, 549 (1947)
/2/ Oohashi et al., Inaudible High-Frequency Sounds Affect Brain Activity: Hypersonic Effect, Journal of Neurophysiology Published 1 June 2000 Vol. 83 no. 6, 3548-3558
  • Last Edit: 10 April, 2017, 08:15:45 AM by Jakob1863

  • Porcus
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #10
First of all, if talking about pseudoscience or scientific reasoning/rigor at all, is archimago´s writing really on par?

Archimago's posting is sufficiently up-to-date to know research that has refuted your reference (Oohashi). So if you want to update yourself, that is likely a good place for you to start.
Have a look at Ashihara and Kiryu (Acta Acustica 89 (2003) no.3 pp. 540ff also known in an earlier AES convention paper version): the Oohashi result is due to intermodulation distortion. It was not the > 20 kHz content, it is the < 20 kHz content. 
Now, Ashihara's team themselves do routinely take care of IM distortion in their own research.  Still they do indeed find that certain subjects can detect 24 to 28 kHz - at ear-splitting volumes, that is. The results from one of those papers (none in that particular study caught the 28 kHz tone) is reproduced in Archimago's blogpost.


There are at least four different topics to discuss:
Brief answers to those: 1: Sure, 2: S(h)ure, 3: depends on equipment, and 4: not shown in actual music, but yes, for certain subjects  > 20 kHz can be detected if played back at levels so loud that the signal would destroy their ability to hear those tones, if the subjects were to actually listen to (hypothetical) music with that content. Whew, too long sentence.


  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #11
4.) do human listeners prefer or at least notice if music does contains energy above 20kHz (during the original event or the reproduction)?

After all these years that is still hard to answer.
No Jakob2. Only yours and other 70yr old elite aural athlete audiophile fantasies.

There were several follow up studies, the most recent one published in 2017, done by various experimenters and in every case some confirmation for those hypersonic effects was found.
Links, not innuendo.

I do remember that pioneer introduced their socalled music-link digital filter in CD players during the 90s because listeners would prefer the reproduction if some content above 20kHz is included.  (afair according to pioneer based on listening tests)
Links, not innuendo.
Loudspeaker manufacturer

  • Jakob1863
  • [*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #12
<snip>
Archimago's posting is sufficiently up-to-date to know research that has refuted your reference (Oohashi). So if you want to update yourself, that is likely a good place for you to start.

May i politely ask that you don´t just trust Archimago´s writing but read the papers by yourself?

Quote
Have a look at Ashihara and Kiryu .......... the Oohashi result is due to intermodulation distortion. It was not the > 20 kHz content, it is the < 20 kHz content. 

The latter is your own conclusion from the former but it is not justified. Mainly because Oohashi et al. did use a seperate "super tweeter" that reproduced the "high frequency content" which is perfectly in line with the results from Ashihara and Kiryu. Methodologically spoken it could otherwise only give rise to valid concerns, but to conclude that Oohashi et al. experimental results were due to would not be justified.

Quote
Now, Ashihara's team themselves do routinely take care of IM distortion in their own research.  Still they do indeed find that certain subjects can detect 24 to 28 kHz - at ear-splitting volumes, that is. The results from one of those papers (none in that particular study caught the 28 kHz tone) is reproduced in Archimago's blogpost. <snip>

I´ve mentioned for good reasons that afaik nobody tried to replicate the experiments done by Oohashi et al.
Other experimenters doing vastly different experiments (means using different stimuli, different test protocols, examing different hypothesis) could provide additional valuable informations but that does not help wrt the hypersonic brain effect.

Archimago is of course right in mentioning that you can´t conclude from those experiments which proportion among listeners could benefit while listening to other music.

That he think´s that Oohashi´s (and that of the follow up studies) is too complicated is at least surprising, it is of course more comprehensive - and that seems to be usually a good idea - and simply tries, as said before, to get more objective data beside the psychoacoustical evaluation.
That we know since Craig Bennett´s "dead salmon story" that false positives are a serious problem in medical imaging is true, but again one has to dig deeper into the various papers to evaluate if their results suffer from these problems.

  • Jakob1863
  • [*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #13
<snip>No Jakob2. Only yours and other 70yr old elite aural athlete audiophile fantasies.

That´s ajinfla´s tales of mistery and imagination.... :)

Quote
Links, not innuendo.

As said before, if there´s interest, i´ll provide a more comprehensive reference list.
Are you interested in real information or just in bickering?

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #14
That´s ajinfla´s tales of mistery and imagination.... :)

 i´ll provide a more comprehensive reference list.
Ok, so this is all your imagination again, zero evidence.
Loudspeaker manufacturer

  • Porcus
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #15
<snip>
Archimago's posting is sufficiently up-to-date to know research that has refuted your reference (Oohashi). So if you want to update yourself, that is likely a good place for you to start.

May i politely ask that you don´t just trust Archimago´s writing but read the papers by yourself?

Of course I have, while you were not even aware of their existence.

"for you to start" meant precisely that, since your reference list was obsolete years ago.

  • Jakob1863
  • [*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #16
<snip>Of course I have, while you were not even aware of their existence.
That´s a funny remark......, but if you have, why did you (incorrectly) report that "Oohashi´s results were due to IMDs and that it would follow from Ashihira´s paper?


  • evgenetic
  • [*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #17
As said before, if there´s interest, i´ll provide a more comprehensive reference list.

it would literally take less time to provide the references than write that very sentence. why the beating around the bush?

  • Porcus
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #18
but if you have, why did you (incorrectly) report that "Oohashi´s results were due to IMDs and that it would follow from Ashihira´s paper?

Quote:
"When each ultrasonic component was not mixed with other components and presented through each of different loudspeakers in order to reduce intermodulation distortions, no subject could detect any ultrasounds in both experiments. It was concluded that addition of ultrasounds might affect sound impression by means of some non-linear interaction that might occur in the loudspeakers. It was also indicated that the ultrasounds would be extremely difficult to detect under an appropriate experimental condition."

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #19
it would literally take less time to provide the references than write that very sentence. why the beating around the bush?
"Good evening, ladies and cherms," says an intense young comic in a nearly empty club. "I chust flew in from Berlin and, boy, are my arms tired."
-Jakob2
Loudspeaker manufacturer

  • Jakob1863
  • [*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #20
Correction: pioneer´s invention was called "legato link conversion" (LLC)

@ evgenetic,

even not taken literally, that´s untrue... :)

@ porcus,
Quote:
"When each ultrasonic component was not mixed with other components and presented through each of different loudspeakers in order to reduce intermodulation distortions, no subject could detect any ultrasounds in both experiments. It was concluded that addition of ultrasounds might affect sound impression by means of some non-linear interaction that might occur in the loudspeakers. It was also indicated that the ultrasounds would be extremely difficult to detect under an appropriate experimental condition."

And now please compare that to the experimental setup used by Oohashi et al. and tell me where Oohashi did something comparable to the situation in which detection in Ashiara´s experiment was possible.
Compare the stimuli (duration, level of spectral components) number of loudspeaker used and effects measured.

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #21
Flap, fap, fap
Loudspeaker manufacturer

Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #22
1.) does music exist that contains some energy above 20kHz?

Surely this question is easy to answer with measurement instruments that have a wider range than human hearing. An objective yes or no must be easy to come by. But permit me to be a little more philosophical.

Why wouldn't it? Or to put it another way, wouldn't it be rather strange if all the worlds instruments didn't produce sounds outside of the range of human hearing? The truly difficult thing would be to design and make all instruments so that they never exceeded that bracket of human hearing. Without the aforementioned instruments.

The question is not is it there, it is is it useful. Is the sound that the composer couldn't conceive, the instrument makers and musicians can't hear, useful? Would it even be pleasant? Higher than the highest hiss? I doubt it very much.

And... No, I have not read the paper(s) first hand, but isn't the problem with Oohashi that he looked for what he wanted to find? Isn't that the problem with quite a lot of people?


 
The most important audio cables are the ones in the brain

  • ajinfla
  • [*][*][*][*][*]
Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #23
i´ll provide a more comprehensive reference list.
Crickets
Loudspeaker manufacturer

Re: Do we "need" those >20kHz ultrasonic frequencies for high-fidelity audio?
Reply #24
First of all, if talking about pseudoscience or scientific reasoning/rigor at all, is archimago´s writing really on par?

There are at least four different topics to discuss:
1.) does music exist that contains some energy above 20kHz?

As Boyk´s already cited paper shows, the answer has to be yes

It also is an irrelevant question. The topic is "what is heard?" not "what exists".  I guess you haven't noticed that we live in a world where almost inconceivable amounts of just about anything can be measured, but the lower measured limits are irrelevant to human life.

So, the above question is a obfuscating distraction from the relevant question.  Shame!

Quote
2.) do microphones exist that are able to work sufficiently in the range above 20kHz?

The answer has to be yes, although certainly not every microphone used for recording is able to

Same problem as above, but the overstatement of the actual incidence of recordings with appreciable output > 16 KHz is noticeable if not telling.  The world is full of so-called professional recording microphones whose output is rolling off rapidly, starting as low as 12 KHz.  How many of which mics are used in any particular recording is not well-known, since it is no longer stylish to mention such things in publicity material or liner notes.  Furthermore, high frequency sound is absorbed by passage though air, and is far less likely to be reflected back into microphones by walls and floors.

Quote
3.) what about intermodulation distortion?

A valid concern, but if the microphones are source of IM distortions then we have to do some acoustical filtering between source and microphone to avoid IMD. Otherwise we have to realize that IMDs are already and inseparable imbedded in the recorded material.

Ignores the fact that the human ear produces large, audible amounts of nonlinear distortion all by itself.

Nonlinear disotrtion in the ear and brain

Quote
The discussion about IMD reminds to a quite similar discussion roughly 60 years ago, when Olson tried to figure out if listeners would prefer full range sound or a bandwidth limited version of the "real thing". /1/

Of course high frequency content means higher risk of IMD, but that would only mean to use better equipment that does not suffer from higher IMDs due to extended high frequency content. (if human listener really prefer content including high frequency sound)


4.) do human listeners prefer or at least notice if music does contains energy above 20kHz (during the original event or the reproduction)?

They don't but for a reason that is very strong but infrequently mentioned: masking.  Sounds that are more easily heard are far more likely to mask sounds that are difficult or impossible to hear.   Just about every recording has a very significant roll-oof of high fequency energy starting as low as a few KHz.   Flether amd Munson proved that our ability to hear higher frequencies, particuarly above 10 KHz drops like a stone in water.

Quote
After all these years that is still hard to answer.

Maybe for the true believers who don't seem to understand that the concept of ultrasonic sound is a well-accepted concept, and that ultrasonic means above the audible range.

Quote
Up to a certain degree that reflects the general unwillingness to replicate studies done earlier.

The simplest possible study is, due to advances in modern technology, painfully easy to do. That you obviously haven't done so for yourself, but continue to pontificate on studies that in the end undermine your viewpoint, tells me exactly  how open-minded you really are, which is to say not at all. :-(

Besides the biggest study one can imagine was done about 10 years ago when the record industry released 100's and even 1000's of so-called hi-rez recordings on SACD and DVD-A that were based on low-rez masters. To this day there don't seem to be any golden eared reviewers who are able to man up and admit that they were fooled by their sighted evaluations, and heard no such problems at all either in the day, or until now.

  • Last Edit: 11 April, 2017, 09:55:01 AM by Arnold B. Krueger