Are we just kidding ourselves?

[mayhem13]

First, thank you to everyone who has contributed to the sheer volume of knowledge this thread now contains; it has clearly exceeded my capabilities for absorption/understanding. But I'm surprised that there isn't more dialog as to these pretty straightforward diagrams (which as Jud has mentioned look a lot like some of the pre-ringing we get from certain filters). Are those that deny any difference from hi-res suggesting that the difference in these waveforms are "inaudible?" What about the data in the article that shows the percentage of listeners able to distinguish between them?

 

It almost feels as though much of this is semantics; i.e. there are those who say we can't hear above 20kHz and thus for them the debate ends because "there can be nothing else" and those who can show there are differences but may not be able to explain how our brains actually measure or "hear" those differences and thus cannot convince the doubters.

 

Wouldn't we be better off recognizing that there can be and are waveform differences (that appear to be audible in some form) and then figure out what resolution/filtering methods allow us to recreate the correct waveform?

 

There's a problem with the relevance to reproduced audio. When we, you or I send these 'improved' transients to our speakers, the smearing of the impulse or step response is much worse than what cutoff filters in the digital domain do to the signal. We're dealing with resonating, moving masses here. Some ribbon tweeters exhibit exceptional impulse response as do plasma devices, but since we can't isolate content below the fundamental of the transient to a different drive unit, the HF device used is poluted either way. Liken the discussion to ultra clear sapphire glass window with a screen in front of it.......not as dramatic of course but I think you can understand the point.

[beanbag]

Shortest: http://artsandsciences.sc.edu/sites/default/files/event_files/January_31_2013.pdf Shorter: http://boson.physics.sc.edu/~kunchur/papers/FAQs.pdf Papers: http://boson.physics.sc.edu/~kunchur/papers/Audibility-of-time-misalignment-of-acoustic-signals---Kunchur.pdf http://boson.physics.sc.edu/~kunchur/papers/Probing-the-temporal-resolution-and-bandwidth-of-human-hearing--Kunchur.pdf http://boson.physics.sc.edu/~kunchur/papers/Temporal-resolution-by-bandwidth-restriction--Kunchur.pdf

 

This guy's work is based on having two speakers offset by a small distance, corresponding to a 5us time delay. What this results in is that the shape of the waveform as heard by the listener gets changed as you change the speaker offset. The shape of the waveform can still be adequately captured via "redbook" sampling frequencies, as per esldude's post earlier. I didn't read the papers closely enough to see what the author did to explain why it is simply not a matter of perceived phase shift. In any case, he did not just play two pulses separated by 5us coming from one source.

[mayhem13]

I've never had an opportunity to listen to a system containing super tweeters but I certainly appreciate reports that super tweeters have made significant improvements in the listening experience.

 

Esau

 

Absolutely. If you can successfully integrate a 'super' tweeter into a speaker system without comb filtering or phase overlap, that's going to be one hell of a speaker to listen to. Comb filtering begins when one freq is being played more than 1/4 the wavelength's distance from another. In the example of isolated super tweeters from 10khz and up that distance to an adjacent drivers acoustic center would need to be .339" or less. So we can see that the avoidance of comb filtering is physically impossible today. I would only add the audible effects of combing are far more pronounced and worse than anything being discussed here in this thread.

[mayhem13]

That mechanism may be possible, but there are far more ways that humans can recognize ultrasonic sound, starting with but not limited to mechanical resonance of structures within the local environment.

 

Got an example of a structure that resonates at plus 20khz that humans have been able to perceive?

 

.....or how bout the trumpet that produces freq at or above 20khz?

[Jud]

This guy's work is based on having two speakers offset by a small distance, corresponding to a 5us time delay. What this results in is that the shape of the waveform as heard by the listener gets changed as you change the speaker offset. The shape of the waveform can still be adequately captured via "redbook" sampling frequencies, as per esldude's post earlier. I didn't read the papers closely enough to see what the author did to explain why it is simply not a matter of perceived phase shift. In any case, he did not just play two pulses separated by 5us coming from one source.

 

From the FAQ:

Each experiment had to be carefully thought out and then submitted as a proposal for approval to an Institutional Review Board (IRB), which is also responsible for legally approving the consent forms that must be used. Then optimum equipment, methods, and a multitude of cross checks must be developed (some details are given in the papers). It takes about half a year to conduct each sequence of controlled blind tests; several sequences were conducted. The results, analysis, and conclusions were then carefully considered and discussed with colleagues who are experts in their related inter-disciplinary fields; for this I went in person to various universities and research institutes and met with people in departments of physics, engineering, psychology, neuroscience, music, communications sciences, physiology, and materials science (as noted in the acknowledgements lists at the end of the papers). After that the results and conclusions were presented at conferences of the Acoustical Society of America (ASA), Association of Research in Otolaryngology (ARO), and American Physical Society (APS). Seminars were also given at numerous universities and research/industrial institutions (please see the list on my web site). After each presentation, the audience is free to tear apart the conclusions and ask all possible questions. Eminent people such as presidents of the above mentioned societies and corporations were present at my presentations and engaged in the discussions. After this oral presentation process, written manuscripts were then submitted to journals where more than a dozen referees and editors were involved in the refereeing process.

I await with interest your explanation of fatal flaws in these papers and experiments that made it past academic and professional experts in various fields as well as the editing and peer review process, but are obvious to someone of your keen intelligence.

[Jud]

There's a problem with the relevance to reproduced audio. When we, you or I send these 'improved' transients to our speakers, the smearing of the impulse or step response is much worse than what cutoff filters in the digital domain do to the signal. We're dealing with resonating, moving masses here. Some ribbon tweeters exhibit exceptional impulse response as do plasma devices, but since we can't isolate content below the fundamental of the transient to a different drive unit, the HF device used is poluted either way. Liken the discussion to ultra clear sapphire glass window with a screen in front of it.......not as dramatic of course but I think you can understand the point.

 

Granting this, I'd think we'd still want as accurate a signal as possible to send to the speakers.

[Jud]

This article is mainly relevant for synthesis of sounds, but is not that relevant to audio reproduction, where all the information is already there.

 

 

Can you tell us where this "there" is that contains "all the information"?

[Jud]

budgie smugglers

 

Always appreciate learning a new descriptive phrase.

[Jud]

In that figure I posted, I have no idea if the differences you can see are audible. What is noteworthy is that the listening tests they report are within the confines of standard resolution, so I don't think it tells us one way or the other about how high res audibly differs.

 

The paper about how percetion violates the Fourier limits is more interesting to me ar least, because it suggests a fundamentally different physiological response to impulse detection. If high res realky makes an audible difference, i bet it will be related to that (in addition to to the other more mundane reasons we for the most part seem to agree on).

 

Not sure of this for two reasons: (1) My guess, and that's all it is, is that better filtering (in the case of high res, done at the studio by better equipment offline, rather than in the DAC chip inline in the case of RedBook resolution material) may allow for less ringing, thus less time smear and perhaps audibly more realistic attack transients. (2) High res is also subject to the Fourier limits; what the research appears to me to show is that the Fourier uncertainty limit (and thus the compromises that must be made between time domain and frequency domain performance in any filter) is within the compass of human audio perception.

[beanbag]

Can you tell us where this "there" is that contains "all the information"?

 

Yes, all the information is contained in the sampled waveform within the analysis window.

If there are N samples, then you get N/2 fourier and phase components to reconstruct the signal.

 

That article uses only about 1/5 of the fourier components to reconstruct the waveform, and then says there is pre-ringing or echo or whatever. So it hardly makes the case that fourier reconstruction is "bad" at reproducing transients, since this isn't how audio reproduction works (except MP3 stuff).

[Jud]

Yes, all the information is contained in the sampled waveform within the analysis window.

If there are N samples, then you get N/2 fourier and phase components to reconstruct the signal.

"[A]ll the information is contained in the sampled waveform...." Within the waveform, within the samples, or both?

 

That article uses only about 1/5 of the fourier components to reconstruct the waveform, and then says there is pre-ringing or echo or whatever. So it hardly makes the case that fourier reconstruction is "bad" at reproducing transients, since this isn't how audio reproduction works (except MP3 stuff).

So digital audio reproduction is not subject to ringing (the Gibbs phenomenon)?

[beanbag]

"[A]ll the information is contained in the sampled waveform...." Within the waveform, within the samples, or both?

 

So digital audio reproduction is not subject to ringing (the Gibbs phenomenon)?

 

Are you asking in context to that article, or in general?

[Jud]

Are you asking in context to that article, or in general?

 

In general.

[Fokus]

In particular, it appears from some academic research that (a) transients/impulses are processed in a different part of the brain than tones,

 

And the relevance is? Hearing/listening is done all over the brain, with various sections of it performing various tasks.

 

and (b) humans may be able to detect transients with a faster rise time than a 20kHz sine wave.

 

Of course. But that is nothing special. Ask yourself: what are the properties of a 'transients with a faster rise time than a 20kHz sine'.

 

- Fourier analysis and Nyquist work on harmonic signals.

 

That is blatantly wrong. There is no such constraint. We are talking about Fourier transforms, not Fourier series.

[Paul R]

I guess to me the question is though in theory reproduction can be perfect at 16/44.1, it is sure easier to come closer to that perfect reproduction when the sample rate and bit depth is higher. It is not so much theory, as a real world application. (shrug) I don't understand why someone would think that all reproductions are perfect just because it is theoretically possible to do so. Most DACs, in fact, oversample to make the conversion easier and less faulty.

 

-Paul

 

 

In that figure I posted, I have no idea if the differences you can see are audible. What is noteworthy is that the listening tests they report are within the confines of standard resolution, so I don't think it tells us one way or the other about how high res audibly differs.

 

The paper about how percetion violates the Fourier limits is more interesting to me ar least, because it suggests a fundamentally different physiological response to impulse detection. If high res realky makes an audible difference, i bet it will be related to that (in addition to to the other more mundane reasons we for the most part seem to agree on).

[mav52]

Some of yall spend so much time worrying about this "stuff", just enjoy the music anyway you can get it. The reason I say that, a good friend of mine of 35 year's, well his wife just passed away Sunday, they were married for 48 years. Both were music lovers, both musicians and singers. They had so much more to do in life together. He told me this morning, you know we worry about a lot of the little things in life and never consider what will be if those bigger things are gone, enjoy the moments and enjoy the music together.

[Jud]

 

That is blatantly wrong. There is no such constraint. We are talking about Fourier transforms, not Fourier series.

 

OK, obviously something I need to understand more about. What's all the stuff in the various journal papers about the great difficulty of modeling non-sinusoidal sounds?

[gmgraves]

i was just pointing out we live in an environment full of ultrasound every day, and except in massive quantity, it is no more like UV exposure than eggs are to Apples.

 

You are concentrating too much on the wrong part of the analogy, I think. That's possibly my fault, but on the other hand, I'm sure you get my point: We can't see ultraviolet light just as we can't hear ultrasonic frequencies. They both might have an effect on parts of the human body (UV gives one a sun-tan/sun-burn, ultra-sonics may be detectable through skin or bone conduction), but neither of these seem to be detectable by the organs of the body dedicated to seeing light or hearing sound.

 

But full range sound is one significant difference between live music and recorded music. If not the ultrasound, then what parts are missing in reproduced sound? Also note, that THD and other factors can come into play.

 

What parts are missing? Dynamic range, lack of distortion, flat frequency response, a truly coherent sound-stage, omnidirectional radiation pattern, etc. But the thing that tells me that ultrasonics is NOT one of the primary differences between live and "canned" music, is the fact that in spite of ultra-sonics being hyper-directional one can be walking down the street and pass an open door to bar or night club and tell instantly that there is live music playing inside. No signs are needed, no one has to announce it, one just knows! It's very doubtful that any ultrasonic frequencies can turn corners to make it out the front door. I've experienced this phenomenon many times (as I'm sure many here have as well). The best way to dramatically experience it is to walk down Bourbon Street at night in New Orleans. this place has live jazz, this place possibly has live music, but I'm hearing sound re-enforcement equipment (yeccchh!) but this next place is playing music out of a juke box. It's often startling how apparent live music is under these circumstances.

 

 

I am not so sure that humans being sensitive to ultrasonics is at all like getting sunburn from a CRT.mat least so long as the CRt vontinues to scan!

 

Again, the part about UV giving a sunburn was a bit of hyperbole on my part, done to make a point and you're right. Our video technology does not produce enough UV to harm us. On the other hand, the part about humans being unable to see UV in the same way (and for the same reason) that we cannot hear ultrasonics (that our sensory apparatus is blind/deaf to these wavelengths) is valid, I think.

[Jud]

And the relevance is? Hearing/listening is done all over the brain, with various sections of it performing various tasks.

 

Of course. But that is nothing special. Ask yourself: what are the properties of a 'transients with a faster rise time than a 20kHz sine'.

 

 

The first relates to the second, and in my mind, further relates to the question of adequate sampling rate. Can transients with a faster rise time than a 20kHz sine wave (or for the sake of exactness, faster than a 22.05kHz sine wave) be reconstructed adequately using a sample rate of 44.1kHz? (Though people seem to be interested in this as a question unto itself, let me reiterate that I tend to think of Redbook vs. hi res not as a question of sample rates but of where the filtering can best be done. The ADC sample rate will be whatever it is, the sample rate in the great majority of DAC chips prior to sigma-delta modulation will be 352.8 or 384kHz, and what's the best way to get from one to the other?)

[sdolezalek]

There's a problem with the relevance to reproduced audio. When we, you or I send these 'improved' transients to our speakers, the smearing of the impulse or step response is much worse than what cutoff filters in the digital domain do to the signal. We're dealing with resonating, moving masses here. Some ribbon tweeters exhibit exceptional impulse response as do plasma devices, but since we can't isolate content below the fundamental of the transient to a different drive unit, the HF device used is poluted either way. Liken the discussion to ultra clear sapphire glass window with a screen in front of it.......not as dramatic of course but I think you can understand the point.

 

Fair comment. About a decade ago we did a lot of work in looking at the impulse responses of individual drivers and built digital correction algorithms that attempted to compensate for the waveform distortion caused by the mass of the magnet and the cone. Although we could show that the resulting waveform was much closer to the original signal than the unadjusted waveform, to our great disappointment very few listeners were able to tell the difference even though the effective distortion levels were much higher than what we measured from all the other equipment in the signal chain.

[monteverdi]

The bone conduction ultrasonics becomes 'audible' when resonances in the skull/brain cause pressure fluctuations in the blood vessels of the inner ear at frequencies below 20 khz. From memory I think the sweet spot is 13-15 khz fluctuations in the blood vessels with the bone conduction sound about an octave higher. The same effect out in the wild would have to be at extraordinarily loud levels without bone conduction. Pretty much it isn't happening that way.

 

Also one of the papers concluded that bone conduction can not discriminate different frequencies. Apparently it only allows to be aware of the presence of frequencies above 20kHz. But without pitch recognition I do not see the relevance for listening to music.

[wgscott]

OK, obviously something I need to understand more about. What's all the stuff in the various journal papers about the great difficulty of modeling non-sinusoidal sounds?

 

The main point is that you can express any periodic function as a discrete sum over sin and cos components, and you can express any (non-pathological) function as a continuous summation (i.e., integration) over of sin and cos components.

[esldude]

Not sure of this for two reasons: (1) My guess, and that's all it is, is that better filtering (in the case of high res, done at the studio by better equipment offline, rather than in the DAC chip inline in the case of RedBook resolution material) may allow for less ringing, thus less time smear and perhaps audibly more realistic attack transients. (2) High res is also subject to the Fourier limits; what the research appears to me to show is that the Fourier uncertainty limit (and thus the compromises that must be made between time domain and frequency domain performance in any filter) is within the compass of human audio perception.

 

It worth pointing out in the research about human hearing exceeding the Fourier limit that the best single result by a single human subject was 3 milliseconds timing discrimination. 16/44 has no problem with such an interval. Further the reason human hearing can exceed the Fourier limit is thought by the researchers to be from non-linearity in the hearing system. This Fourier limit is about discriminating frequency and time of signals, not pure time or pure frequency.

 

So while a linear system like any rate of PCM would be subject to Fourier limits it doesn't mean what is being implied. Which is an inability of PCM to portray timing equal to or surpassing that of human hearing with normal 44/16 redbook format, and further implying that higher sample rates help with this issue.

[esldude]

Also one of the papers concluded that bone conduction can not discriminate different frequencies. Apparently it only allows to be aware of the presence of frequencies above 20kHz. But without pitch recognition I do not see the relevance for listening to music.

 

Yes, you are correct. In fact most of the research is from using this non-pitch specific cognition of bone conducted ultrasonics to combat tinnitus. Somewhat masking and reducing the undesired effects of such a problem.

[Jud]

The main point is that you can express any periodic function as a discrete sum over sin and cos components, and you can express any (non-pathological) function as a continuous summation (i.e., integration) over of sin and cos components.

 

So is (to use the description from the IEEE article abstract) "highly nonstationary inharmonic behavior" getting close to "pathological"? I think I'm correct in saying at least a good portion is non-periodic.