Hypersonic effect: high frequency spectrum listening experience

[DeVerm]

This thread is a continuation of off-topic posts that started with post #96 on this page: http://forums.slimdevices.com/showth...=53345&page=10

It's about the research done in Japan that was published in 2000 of which you can find a copy here: http://jn.physiology.org/cgi/content/full/83/6/3548
Better PDF copies that include more tables with data can be found on the web.

So, here we pick it up:

You regard an internet search for copyright-constrained ABX samples as evidence that "most" ABX testing uses short samples, and from there conclude that the ABX protocol is invalid? Don't be silly.

Okay, I'll drop that because I seem incapable of explaining the difference between the ABX protocol (okay) and the wide-spread implementation (not okay) of it. It's not the core-issue here.

Of course it's not! Those are two tests out of thousands - and you're trying to conclude something about MOST abx tests?

The first of these two studies (Muraoka et al. 1978) is more important in our current context than all the others because the current Redbook specification is a result of it. Also, this study used the CCIR recommendation of 15-20 second samples and 0.5 second interval between samples. Also, the test subjects did a subjective listening test and noted their findings in a questionnaire; all input for the conclusions was taken from these questionnaires and the subjects did not have their brains scanned. When I say that it was flawed, it's because it turned out that by using that procedure, it is not possible to find any evidence for perception of > 20 kHz harmonics by human listeners. They checked that by repeating it with the same music-samples and sound-equipment ("presentation system")used for their new method and the results were in agreement with these of Muraoka et al. (1978).

PART of their results agree (the part that show that the HFS alone are inaudible), and PART disagree.

No, that "disagreed part" you mention was never tested before. The subjects still can't hear the HFS part even when it's played together with the LFS part... but --their brains react to it--. The scope of previous tests never included that possibility and thus did not monitor the brains of the subjects. When working from two different scopes like that, you can't say that there is disagreement as the first test didn't include this part. You can say that the first test missed it.

I'll try one more time. Everyone seems to agree that HFS alone are inaudible based on PET brain scans etc.

Ack.

But these guys find that HFS+LFS is different from just LFS. So there is something very bizarre and non-linear going on if they are right.

Nack. Why is this bizarre? The HFS component is what is in the original sound as from the instruments: high freq. harmonics. When you move the 20 kHz line down to lets say 12 kHz and take a recording of an acoustic instrument without it's harmonics above the 12 kHz "everyone" hears the difference between HFS+LFS vs just LFS. When you play just the HFS, it might even be totally un-recognizable as the instrument. In other words: high harmonics add to the fundamental+low harmonics but are utterly senseless on their own. That's why a very expensive violin sounds better than a plastic toy.

The next step up is to make the LFS < 20 kHz and the HFS > 20 kHz and the interesting thing is that humans can't hear the HFS anymore... but their brains register the presence of it anyway! If you call that bizarre it is because you can't let go of the notion that only your ears feed your brain with information when listening to music. Other research like mentioned in other threads has already showed that this is not the case, like even memories of hearing this song or smell or seeing how someone else reacts to the music are all factors for what happens in your brain while listening and thus change the listening experience. These are established facts. Also, very low freq's you will have a hard time of hearing them but you feel them and that changes your listening experience too. So, somehow, no-one knows yet exactly how, very high harmonics are sensed by humans as demonstrated in this study. If you can only accept that after they prove exactly --how-- that is done, that's fine with me, but we all sense it while you're waiting for that proof. (I know, I do it again but I just can't help myself, sorry ;-)

My point was that we have no way of knowing whether that bizarre non-linear thing is in their equipment or in people's heads.

That is a question that must be answered for all research done. The established method is by using multiple and totally different methods of measuring. The primary method they used here was the EEG scan, in the alpha range. These are electrodes that measure brain-patterns in the couple Hz range, like 6 Hz or so. But you are right, the EEG scanner or electrodes could have interference by the high frequency sounds in the room. It is very unlikely because the sensors can't pickup those freq's and while measuring with just the HFS, it didn't show anything. They could even make a measurement without placing the electrodes on the subjects head to prove accurate EEG recordings... but they did something better: a second and totally different measurement of blood-flow in their brains using a PET scanner. This equipment detects radio-active radiation from the brain after injecting the subject with a radio-active solution. There is no way that playing music can interfere with this as there is no radio-activity present in the sound-waves, let alone that it is interfered in exactly the same way as it could do with the EEG method. Both methods alone are very unlikely to have interference from playing music but when both measurements show the strong correlation that they did in this case, you eliminate the chance of flawed measurements completely. This is standard and accepted practice.

To summarize: we know that neither brains nor gear respond to HFS alone, but that brains+gear respond to HFS+LFS differently than to LFS alone. But we don't - and can't - know whether that difference is due to brains or due to gear. The fact that gear doesn't respond to HFS alone is irrelevant, because neither do brains!

Would you be happy when they add a third method, like MRI scan? When do you start believing measuring instruments? The scientific community believes measurements when they can be duplicated using a completely different method of measuring and that is good enough for me and most people. It is why they use EEG in hospitals and work with it's results. They only add a PET or MRI scan when they think the results are outside their expectations but use just 1 method otherwise. Also, note that the study was done by many medical professionals, not just a bunch of audio technicians. They know how to use EEG and PET equipment and have done it many times before on patients etc.

cheers,
Nick.

[sebp]

You said that :

The document clearly states that Pioneer is the manufacturer and not Tsutomu Oohashi. Oohashi works for 1) Department of KANSEI Brain Science, ATR Human Information Processing Research Laboratories, Kyoto; and 2) Department of Network Science, Chiba Institute of Technology, Narashino. So he's a researcher that developed a speaker for Pioneer which is very normal and done by many researchers when a manufacturer doesn't have the knowledge in-house.

If he designed the speaker he was paid by Pioneer, and may well receive royalties on sales. This article will promote sales of speakers with supertweeters. That's about as clear a conflict of interest as you can get.

I read that :

Then, LFCs and HFCs were separately amplified with P-800 and P-300L power amplifiers (Accuphase, Yokohama, Japan), respectively, and presented through a speaker system consisting of twin cone-type woofers and a horn-type tweeter for the LFCs and a dome-type super tweeter with a diamond diaphragm for the HFCs. The speaker system was designed by one of the authors (T. Oohashi) and manufactured by Pioneer Co., Ltd. (Tokyo, Japan).

Could it be simply possible that Mr Oohashi, for this experiment, asked Pioneer to manufacture speakers according to his specs?

[DeVerm]

Could it be simply possible that Mr Oohashi, for this experiment, asked Pioneer to manufacture speakers according to his specs?

I even suspect that these speakers were never sold at all but just manufactured for this test. I see that Pioneer does sell speakers that can produce up to 100 kHz sound but none have two tweeters as described in the paper. This diamond diaphragm can't produce lower high-range (if it would, there wouldn't have been two tweeters) so it's not in the line of speakers sold by Pioneer today.

cheers,
Nick.

[DCtoDaylight]

Hope this moves over to the new thread properly...

Medical ABX tests routinely run for months and sometimes years, in order to properly evaluate the results. It may not be convenient to do that with audio gear, but I personally believe that's what's required.

Explain me how to do that? you can't sit and listen to a sample for months or years...??

Why not? Two full, album length recordings, each at different sample rates. You are given two years to decide which is better... I didn't say it was easy, what I said was it's possible.

I see too many cases of people claiming ABX testing is flawed or can't reveal the truth, when in fact, it isn't ABX testing that's at fault, but rather it's a specific implementation that's at fault.

Cheers, Dave

[opaqueice]

No, that "disagreed part" you mention was never tested before. The subjects still can't hear the HFS part even when it's played together with the LFS part... but --their brains react to it--. The scope of previous tests never included that possibility and thus did not monitor the brains of the subjects. When working from two different scopes like that, you can't say that there is disagreement as the first test didn't include this part. You can say that the first test missed it.

That's just... wrong. You're directly contradicting the authors of this paper:

Explanation of the discrepancy between the present and previous studies

The fact that we used an entire piece of natural music lasting 200 s as sound stimuli instead of short fragments of sounds might explain the discrepancy between our findings and those of previous studies carried out around 1980 to determine the format for digital audio CDs (e.g., Muraoka et al. 1978; Plenge et al. 1979), which concluded that the presence of sounds containing a frequency range above 15 kHz was not recognized as making a difference in sound quality.

Those studies (probably among many others) showed that people cannot distinguish between sounds including high frequencies and sounds not including them. That directly contradicts the findings of this study, and the method (on the question/response part of this research) was essentially identical AFAIK. The only significant difference is the length of the sample.

Which brings up another problem with this. If the authors' proposed explanation is correct (that the effects only manifest themselves over time periods of 10s of seconds), it's perfectly possible that any presence of HFS (regardless of whether it's harmonically related to the music) could activate these beneficial effects. In other words...

Hypothesis: people feel strange in environments (such as the damped soundbooths this experiment probably used) with zero HFS present. Adding HFS to music in an otherwise completely silent environment "reassures" the brain that all is well, whereas playing only LFS leaves this strange feeling. In a normal living-room type listening environment with ambient HFS reproducing the HF content of music would have no effect, because the background HFS field would "reassure" the brain on its own.

If that hypothesis is correct - and as far as I know it's consistent with their data, even accepted at face value - these findings would have no relevance whatsoever for home audio.

Nack. Why is this bizarre?

Fact: neither people's brains nor conscious minds react to HFS alone.

Claimed fact: people's brains and responses are different when exposed to HFS+LFS versus LFS alone.

That's very weird.

It's the kind of thing that is out there enough that it could have many potential explanations (for example the one I gave above). Once you allow for possibilities like that, everything should be questioned - for example, I repeat, how do you establish that the measuring gear you're using doesn't exhibit this kind of non-linear response? You really can't until you have a self-consistent theory for where the effect is coming from, and so far there isn't one.

[Themis]

Could it be simply possible that Mr Oohashi, for this experiment, asked Pioneer to manufacture speakers according to his specs?

True. But after this experiment, several manufacturers started producing super tweeters. It's quite common nowadays speakers (and headphones) delivering 30-50kHz.

[DeVerm]

Why not? Two full, album length recordings, each at different sample rates. You are given two years to decide which is better... I didn't say it was easy, what I said was it's possible.

I see too many cases of people claiming ABX testing is flawed or can't reveal the truth, when in fact, it isn't ABX testing that's at fault, but rather it's a specific implementation that's at fault.

I agree with most you write, but your example would loose the "blind" part of it as you can see which version has the higher sample-rate. Also, for abx you would need three recordings of which two are the same... if I understand it all correctly.

But again, I agree and one could be given 3 black boxes labeled A, B and X with only on/off switch and analog output and let you play with it for as long as you want. But, two years? after 2 years you would be listening to obsolete stuff!

cheers,
Nick.

[Phil Leigh]

I agree with most you write, but your example would loose the "blind" part of it as you can see which version has the higher sample-rate. Also, for abx you would need three recordings of which two are the same... if I understand it all correctly.

But again, I agree and one could be given 3 black boxes labeled A, B and X with only on/off switch and analog output and let you play with it for as long as you want. But, two years? after 2 years you would be listening to obsolete stuff!

cheers,
Nick.

Nah - I'd probably still be listening to "Wish You Were Here" )

[lanierb]

I'm jumping in late here, but is there a theory about how exactly these high frequency harmonics make any difference to a human whose ears can't render them? Clearly you can't feel them either as you can with low frequency. Is there meant to be some other "sense" that picks them up? Sounds crazy to me for this simple reason.

Edit: OK I just found the other thread where at least one person raised the issue already. Answer: there is no serious explanation, just wild speculation.

[sebp]

Since sound is just air vibrating, I can't see no reason why your whole body - and not only your eardrums - could not act as a receptor.
I believe skull bone is being used as a resonator for helping hearing-impaired people.

Ever felt infra-basses in your stomach during a concert?
It's not about hearing low or high frequencies, it's all about feeling them (consciously or not).

[DeVerm]

That's just... wrong. You're directly contradicting the authors of this paper:

Well, it seems that you insist that the methods used were the same as in earlier studies so it's no use to keep arguing about this. But, for the record, I state that studies by Muraoka et al. (1978) and Plenge et al. (1979) did use questionnaires only and not EEG or PET scans of the brain. The paper states that also. You disagree?
Furthermore: this study used 200s samples (complete song) and 10s interval. I state (as does the paper) that the studies by Muraoka et al. (1978) and Plenge et al. (1979) did use 15-20s samples and 0.5s intervals. You disagree? (you also never seem to answer the questions I post so it's impossible to work down to the core of your disagreement... you evade.)

Lastly, they did studies using EEG before (Oohashi et al. 1994) and the findings are in agreement with that study.

Those studies (probably among many others) showed that people cannot distinguish between sounds including high frequencies and sounds not including them. That directly contradicts the findings of this study, and the method (on the question/response part of this research) was essentially identical AFAIK. The only significant difference is the length of the sample.

You repeat again. Your AFAIK is fake because you know that "independent of a subjective evaluation of sound quality" EEG and PET measurements are differences too. I'll quote it here:

"In our EEG and PET experiments, we focused on physiological brain responses and objectively evaluated the effect of the combination of audible sounds and inaudible HFCs on brain activity, independent of a subjective evaluation of sound quality."

You also ignore the third part of the current study where they followed the same procedure as in the studies by Muraoka et al. (1978) and Plenge et al. (1979) and found the same results as they did in the 70's.

Hypothesis: people feel strange in environments (such as the damped soundbooths this experiment probably used) with zero HFS present. Adding HFS to music in an otherwise completely silent environment "reassures" the brain that all is well, whereas playing only LFS leaves this strange feeling. In a normal living-room type listening environment with ambient HFS reproducing the HF content of music would have no effect, because the background HFS field would "reassure" the brain on its own.

If that hypothesis is correct - and as far as I know it's consistent with their data, even accepted at face value - these findings would have no relevance whatsoever for home audio.

You read the paper so you know that they describe the rooms where the tests were performed incl. the decorations, paintings on the wall and even the view from the window. But now you say that they "probably used damped soundbooths" which isn't trolling because I know you would never do that... but it feels like it is anyway. You also say that normal listening environments have ambient HF but I wonder how you know that or what it's source might be. Especially how that would be at the dB levels of playing music. The only info I find is that HF is present in natural environments like the rainforest, but I assume that audiophiles prefer a non-natural environment like a house (or boat).

Fact: neither people's brains nor conscious minds react to HFS alone.

Claimed fact: people's brains and responses are different when exposed to HFS+LFS versus LFS alone.

It is not claimed anywhere that HF sound changes the brain. But it is an demonstrated fact that full range music results in a significant larger brain-response are shown by both EEG and PET scans as compared to the same music without the HF components. You can't accept that.

But you --can-- accept that listening experience changes when you consume for example alcohol or have a good(!) smoke. Or that it changes in the presence of other fans of the song etc. In these cases, do you think that these factors change your ears? It is established that it doesn't change your ears but invokes other brain activity which, in combination with the music you hear, changes your listening experience. It's no different with this study when you replace the alcohol/smoke/fans with HF audio components.

Now the BIG question: do you hear a difference between 44 and 48/96 kHz sampled music? Many, many audiophiles do and indeed, releasing these HD formats more and more (now 5000 SACD titles and 5000 DVD-audio titles) indicate that many prefer it. Why?

I repeat, how do you establish that the measuring gear you're using doesn't exhibit this kind of non-linear response? You really can't until you have a self-consistent theory for where the effect is coming from, and so far there isn't one.

Again you question the established EEG and PET scans. I wonder what you're gonna do if you would need any of these in a hospital? Question it's linear response with the doctors? I can not for my bone-marrow understand this non-linearity thing as neither measure anything linear related. What is so non-linear about measuring increased blood-flow in ones brain? Believe me, if a PET scan shows increased blood flow in part of your brain, you -have- increased blood flow in that part of your brain; similarly, if an EEG scan shows increased brainwave activity, you -have- increased brain activity (it's the key in defeating lie-detectors). If you don't accept that, it's your loss but I have no further wish trying to convince you of this (I fail obviously) so I'll leave it at that.

On the theory for explaining the difference; how about the one in the paper:

"sound frequencies in the audible range function as a message carrier and frequencies above the audible range, together with those in the audible range, function as a modulator of sound perception through the brain systems, including the reward-generating system."

It will be many many decades or centuries before it is possible to explain how this works, because we know so little about how the brain works and learn so slow. That is no reason to ignore the fact that most, if not all, humans have a more satisfying listening experience when they use HD material and especially audiophiles should explore and enjoy it. Good news for Sean: we all need a transporter!

Like I wrote before, these findings were 8 years ago and we have accepted it and moved on to HD, no matter if you believe it or not. But now you have an explanation of why audiophiles claim that one sample over the other is "more satisfying" and similar wordings.

cheers,
Nick.

[DeVerm]

I'm jumping in late here, but is there a theory about how exactly these high frequency harmonics make any difference to a human whose ears can't render them? Clearly you can't feel them either as you can with low frequency. Is there meant to be some other "sense" that picks them up? Sounds crazy to me for this simple reason.

Edit: OK I just found the other thread where at least one person raised the issue already. Answer: there is no serious explanation, just wild speculation.

Exactly, there is no consensus whatsoever on explaining why this happens. The simple fact that it happens remains and is only disputed by die-hard red book fans of the Mr. O kind ;-) Another fact is that one can refuse to believe something forever and one can always find new objections and reasons for this not-believing-it. Some people still don't believe that man walked on the moon and no one can convince them that we did. But luckily, that doesn't stop progress too much so we now have HD music for sale and can play en enjoy it and send nifty little robots to Mars etc.

cheers,
Nick.

[darrenyeats]

The simple fact that it happens remains and is only disputed by die-hard red book fans

Whoa. Even die-hard red book fans, as you put it, have always known it is possible to hear the difference between red book and hi-rez under certain conditions.. In this sense this study isn't really news. The way I see it using "the gamelan music of Bali, which is extremely rich in HFCs with a nonstationary structure" is a bit like playing silence very loudly. It's trying to prove a point...but not a point I'm interested in.

What about some nice classical piece, aren't there enough instruments of sufficient frequency range in there? Picking a genre, say classical, because it has HF content is as far as I'm willing to go. Any more picky than that and you're rigging the game so much I don't care about the outcome. And before you say it I rate Gamelan music a 'genre' as much as I would silence...or square waves. :^)
Darren

[Themis]

Part of this study refers to the redbook upper frequency cut.
Also, as I understand, part of this study refers to the ABX methodology.

In my opinion, ABX tests and the statistical methodology that accompanies them are NOT DIRECTLY questioned in this experiment.

Nevertheless, ABX tests are used to try to establish a difference or to establish the absence of a difference (similitude) among the way two reproduction machines can reproduce music. So we use sound samples as means of conducting these tests, then we proceed to the conclusion (are they similar or not).

We knew before this experiment that :
- Frequency samples cannot be used to distinguish two reproduction machines using ABX tests.
Explanation: A music message is a sum of primary (for instance: sinusoidal) sounds. This is similar to the way a D/A conversion works. One could think that if we can distinguish a difference using two frequency samples, then we can conclude that we can -logically- distinguish the resulting (as a sum) music.
In fact, several tests (I participated myself in one of them) proved that we CANNOT distinguish two frequency-based sound samples reproduced by two different machines, even if we could distinguish (using the same ABX tests) these machines using normal (time-based) music samples.
Why ? Simply because a human ear is NOT a linear measurement instrument. Not really a big discovery, some may say. Exactly: Oscilloscopes and other measurement instruments are much better than our ears.
Conclusion: Never use frequency-based (primary) sounds to proceed to ABX tests: The results are irrelevant, the tests are void.

But we have always thought that:
- Duration samples (complex, but of a small duration) could be used to distinguish two reproduction machines using ABX tests.
This experiment concludes that duration samples cannot be used to proceed to ABX tests, because the results may be irrelevant.
Well what then ? What does it mean may be ? They are, or they are not ?
In fact, as ABX tests use mathematical (statistical) methodology, the range rule apply: the measurement means MUST be absolute. No maybe is accepted, or the methodology is broken. Statistics need facts to work, not approximations.
So, in fact, as this experiment establishes an approximation (or even the not measurable possibility of an approximation, if you are optimistic), the samples cannot be used anymore. Simple mathematics.

So what ? No problem: It's enough to use longer, complex samples to validate the ABX tests, that's all. No problem, as I stated.
On the other hand: any ABX test that doesn't use long, complex samples is void. We can't manipulate maths as we want, pity.

[opaqueice]

Well, it seems that you insist that the methods used were the same as in earlier studies so it's no use to keep arguing about this. But, for the record, I state that studies by Muraoka et al. (1978) and Plenge et al. (1979) did use questionnaires only and not EEG or PET scans of the brain. The paper states that also. You disagree?

No, that's exactly what I said (several times already). The questionnaire/listener response part of this study was more or less identical - other than the length of the samples - to those previous papers. Its results contradict those of the previous studies.

(you also never seem to answer the questions I post so it's impossible to work down to the core of your disagreement... you evade.)

No, you just don't read my posts.

You repeat again. Your AFAIK is fake because you know that "independent of a subjective evaluation of sound quality" EEG and PET measurements are differences too.

I said "on the question/response part of the study". There is no point in continuing this if you don't even read what I write and instead attack something from your own imagination (as you have done repeatedly).

But now you say that they "probably used damped soundbooths" which isn't trolling because I know you would never do that... but it feels like it is anyway.

That was for the EEG part of the experiment. As far as I can see they never describe the conditions under which the questionairre part was conducted.

You also say that normal listening environments have ambient HF but I wonder how you know that or what it's source might be. Especially how that would be at the dB levels of playing music. The only info I find is that HF is present in natural environments like the rainforest, but I assume that audiophiles prefer a non-natural environment like a house (or boat).

All natural sounds have harmonics. Precisely how much is present will be a function of the specific environment. But just as an example, an NTSC TV screen produces a sound around 16kHz (which I can hear, by the way) plus harmonics. Computer monitors also produce HF sounds, as do florescent lights, and probably lots of other household electronics.

Now the BIG question: do you hear a difference between 44 and 48/96 kHz sampled music? Many, many audiophiles do and indeed, releasing these HD formats more and more (now 5000 SACD titles and 5000 DVD-audio titles) indicate that many prefer it. Why?

There is no evidence for that assertion. Time and time again audiophiles have made such claims, only to fail completely when asked to distinguish blind. A recent case was the failure of hundreds of subjects tested over two years to distinguish between 16/44 and SACD. That study was much more directly relevant to home audio than this one. Not one subject succeeded.

Again you question the established EEG and PET scans.

I'm simply applying precisely the same convoluted logic to the instruments the experimenters applied to human hearing.

Like I wrote before, these findings were 8 years ago and we have accepted it and moved on to HD, no matter if you believe it or not.

What world are you living in? In mine, hi-res audio formats are dead (albeit still twitching occasionally).