http://www.physics.sc.edu/kunchur/Acoustics-papers.htm (http://www.physics.sc.edu/kunchur/Acoustics-papers.htm)
http://www.physics.sc.edu/kunchur/papers/FAQs.pdf (http://www.physics.sc.edu/kunchur/papers/FAQs.pdf)
I'm not sure wether they have actually tested the impact of high frequencies on human hearing or just the impact of a single-capactitor low pass filter.
Nice to see this topic here where it can be discussed calmly and rationally.
--Ethan
from the link:
Our recent behavioral studies on human subjects proved that humans can discern timing alterations on a 5 microsecond time scale, indicating that that digital sampling rates used in common consumer audio (such as CD) are insufficient for fully preserving transparency.
Exactly how does some air vibration < 18kHz only last 5 microseconds?
Also,
.000005 = 200 KiloHertz
Does this mean we need > 400kHz sampling rates?
Does this mean we need > 400kHz sampling rates?
Well, I guess what we can conclude from Prof. Kunchur's research is that indeed, you might need 400 kHz to digitally represent
a 7-kHz square wave transparently. I'm not listening to such square waves in my free time very often, though.
Would be curious to see what vinyl makes out of a "perfect" 7-kHz square wave.
Chris
Well, I guess what we can conclude from Prof. Kunchur's research is that indeed, you might need 400 kHz to digitally represent a 7-kHz square wave transparently. I'm not listening to such square waves in my free time very often, though.
Well, 400kHz aside, if he was anywhere vicinity of being right - what I'm not willing to swallow, yet - the long time objectivist argument would be broken:
- Humans can't hear anything above 20kHz.
- 44.1kHz sample rate is enough to cover all that completely according to Nyquist.
- -> The Redbook storage format is completely sufficient for transparency.
Well after reading through the first 3 PDF's and the FAQ, I surmise (from my novice knowledge of digital audio concepts) that his main point is centered on arrival times / phase differences. His blind testing groups could identify with good confidence down to about 5 microseconds. Apparently bandwidth restriction (44.1 kHz sampling for example) and loudspeaker placement (within a few millimeters) can each independently introduce timing variances > 5 microseconds that can be blind-test identified. Also there is a mention of two ultrasonic off-phase samples which cause (unwanted) lower sonic harmonics that can be identified as well.
I wonder what kind of design changes and production costs could accommodate typical electronic audio hardware sampling at > 192 kHz or even > 400 kHz? And what software encoding scheme would be required?
I'm still why the usually quite vocal "44.1kHz ought to be enough for anybody"-crowd doesn't take a stand on this...
I have prepared a set of audio files to verify Professor Kunchur's claims for the domain of digital standard vs. high rez audio, that is not subject Kunchur's lowpass circuitry. 7kHz square waves were directly generated into the corresponding output formats. It was quite difficult to get done, even Audition 3 could not generate 7kHz squares without notable artifacts. The results are interesting! You'll need a high end DAC, though.
Use files of equal bit rate for ABX testing! They are normalized to -10db and have short fade-in/-outs applied to prevent transient clicks while looping:
32 bit, 192 kHz:
[attachment=5263:192kHz.wav]
32 bit, 192 kHz upsampled from 44.1kHz (Sox VHQ):
[attachment=5264:192kHz_f..._44_1kHz.wav]
32 bit, 110 kHz, optimized for Benchmark DAC1s:
[attachment=5266:110kHz.wav]
32 bit, 110 kHz (DAC1) upsampled from 44.1kHz (Sox VHQ):
[attachment=5267:110kHz_f..._44_1kHz.wav]
32 bit, 44.1 kHz (for reference only):
[attachment=5265:44_1kHz.wav]
Please re-download! I had accidentally uploaded the wrong set of files.
PS These are 32 bit integer files, which is Sox' default. Some applications (Audition and as reported even Foobar) have trouble playing them. Convert them to 32 bit float if you are affected. I also opted for integer because they are easier to verify with a hex editor.
rpp3po, the 192kHz versions are still different. The 44.1kHz sample matches the upsampled sample, but not the raw 192kHz sample, verified by ears and spectrograms.
Ok, I have re-generated them again and am going to re-upload.
In the meantime, does anyone see any flaws here?
mbp:~ rpp3po$ sox -r 44100 -n 44100.wav synth 5 square 7000 gain -10
mbp:~ rpp3po$ sox -r 192000 -n 192000.wav synth 5 square 7000 gain -10
mbp:~ rpp3po$ sox 44100.wav 44_1kHz.wav fade .010 0 .010
mbp:~ rpp3po$ sox 192000.wav 192kHz.wav fade .010 0 .010
mbp:~ rpp3po$ sox 44_1kHz.wav 192kHz-from-44_1kHz.wav rate -v 192000
PS This (http://sox.cvs.sourceforge.net/viewvc/sox/sox/src/synth.c?revision=1.72&view=markup) is the source code of Sox' square wave generator.
Don't even bother using a square wave generator in an audio editor - in order to ensure that the aliasing is below a 16-bit noise floor, you'd need 65536x oversampling....
Instead, construct the square wave by hand using additive synthesis based on the Fourier series expansion:
Amplitude(n) = 1/n, odd n; 0, even n
Phase(n) = 0
(Or, if you are sure your audio editor uses a technique immune to aliasing issues, like this one, use it.)
Such techniques are well documented - it's quite a shame that so many audio applications (and Dr. Kunchur, and other audiophiles) continue to use bad code.
Yes, I asked myself why Adobe would even include such a broken feature. It is very obviously broken and they should have seen that. Sox' results look fine to me now, though. If not, I'm open for feedback.
Such techniques are well documented - it's quite a shame that so many audio applications (and Dr. Kunchur, and other audiophiles) continue to use bad code.
I thought that Kunchur had used an analog square wave generator?
I only skimmed the paper, but IIRC tried a digital one couldn't get it to work (for unspecified reasons) and then used an analog one. I presumed it was because he had a high end analog synthesizer handy (they're pretty common in labs since people used them for all sorts of stuff in the days before cheap digital DAQs).
Now the WAV files won't load in foobar2000...
Edit: Hacked around with them, got them loading in foobar2000, but all of them have subharmonics well under 7000Hz. At first I thought I was just hearing some weird IMD, but the subharmonics are there.
Edit 2: Synthesized my own versions, using Axon's cited additive synthesis technique:
http://benjamincook.ca/441.wav (http://benjamincook.ca/441.wav) - 44.1kHz square, harmonics at 7k (gain 1) and 21k (gain 1/3).
http://benjamincook.ca/192-441.wav (http://benjamincook.ca/192-441.wav) - 44.1kHz square, harmonics at 7k (gain 1) and 22k (gain 1/3), resampled to 192kHz using sox 441.wav 192-441.wav rate -v 192000
http://benjamincook.ca/192.wav (http://benjamincook.ca/192.wav) - 192kHz square, harmonics at 7k (gain 1), 21k (gain 1/3), 35k (gain 1/5), ..., 91k (gain 1/13)
These really don't look square in any editor, but they should be mathematically-acceptable. I really can't ABX these. It hurts my ears, and I don't have a DAC that handles 192kHz nicely.
Edit 3: For the curious, this is simply the <math.h> sin function, at 32-bit floating-point precision. Fixed some numbers in Edit 2.
Nice to see this topic here where it can be discussed calmly and rationally.
I'm wondering why this thread isn't getting more attention. It should be much more original HA territory than the "Why we need audiophiles" juggernaut.
Kunchur's claims were introduced here at HA by
moi two weeks ago:
http://www.hydrogenaudio.org/forums/index....mp;#entry646398 (http://www.hydrogenaudio.org/forums/index.php?showtopic=47827&st=160&p=646398&#entry646398)
and yes, rpp3po , he is saying the Redbook is broken in terms of transparency.
Hence the uproar on Stereophile's forum, where it's delightful to see what fulsome respect the letters 'PhD' can garner from audiophiles when they really want to believe.
Well, 400kHz aside, if he was anywhere vicinity of being right - what I'm not willing to swallow, yet - the long time objectivist argument would be broken:
- Humans can't hear anything above 20kHz.
- 44.1kHz sample rate is enough to cover all that completely according to Nyquist.
- -> The Redbook storage format is completely sufficient for transparency.
Just as an aside, no, that's not the case.
In order to contain the bandwidth of a signal, you have to filter it. It is possible (i.e. it is done with ridiculous filters which I cheerfully stipulate are not useful in any real sense) that filters might have a slight, tiny effect, maybe, kinda sorta, PERHAPS, at 44.1. Even less likely at 48, and not at all at 64. Nobody has shown this with sensible filters, by which I mean filters that have decent transition bandwidth (i.e. not as tight as humanly possible), ripple, and stopband rejection.
And, of course, if Dr. K's argument is mistaken, that shows nothing, for or against.
These really don't look square in any editor, but they should be mathematically-acceptable. I really can't ABX these. It hurts my ears, and I don't have a DAC that handles 192kHz nicely.
Edit 3: For the curious, this is simply the <math.h> sin function, at 32-bit floating-point precision. Fixed some numbers in Edit 2.
Could anybody enlighten me why a sine function, that supposedly outputs something that is not square (can't check - files are offline right now), should be a better approximation of a square wave than successive sequences of -x,-x,-x,-x,-x,-x,-x,-x,+x,+x,+x,+x,+x,+x,+x,+x values and x being a constant?
Could anybody enlighten me why a sine function
Not a sine, but a sum of sines:
Amplitude(n) = 1/n, odd n; 0, even n
Phase(n) = 0
is 1*sin(f*x) + 1/3*sin(3*f*x) + 1/5*sin(5*f*x) + ... + 1/N*sin(N*f*x), where N*f < 2*pi*Nyquist_frequency.
that supposedly outputs something that is not square (can't check - files are offline right now), should be a better approximation of a square wave than successive sequences of -x,-x,-x,-x,-x,-x,-x,-x,+x,+x,+x,+x,+x,+x,+x,+x values and x being a constant?
Because of aliasing. Take analog square wave and sample it
without lowpassing it at Nyquist freq. You'll get that square digital wave; it contains frequencies below Nyquist limit and aliases of frequencies
above it.
Why not just just create a 7.35kHz square wave "manually", i.e. for 44.1kHz sample rate 6x 32767 followed by 6x -32768 <repeat>; 15x for 110.25kHz and 24x for 176.4kHz?
Could anybody enlighten me why a sine function, that supposedly outputs something that is not square (can't check - files are offline right now), should be a better approximation of a square wave than successive sequences of -x,-x,-x,-x,-x,-x,-x,-x,+x,+x,+x,+x,+x,+x,+x,+x values and x being a constant?
The sum of sins approach is exact for a band limited square wave (since a band limited square wave is by definition the Fourier series of a non-limited square wave truncated at the band limit). Flipping between +/- x and then low pass filtering is only an approximation thats limited by the quality of the filtering applied.
Because that places unacceptable restrictions on the desired wavelength.
OK, so, my paper is taking a little too long to get out the door, so I will provide an executive summary here. I am trying as fast as possible to get it out, but the discussion is about to pass me by, soooo...
- His definition of the power difference at 14Khz, ΔLp(2), is not derived. When one derives it, one discovers that he fails to correctly sum the 14kHz components resulting from the (7+7) and (21-7) terms - in fact, he explicitly treats the calculation of ΔLp(2) as a "peak" level difference, when in fact the summed component quite obviously has a constant level. (That Dr. Kunchur makes such a basic trig error is profoundly disturbing to me.)
- His definition of ΔLp(2) also uses the theoretical values of Δφ instead of the values he already documented in his own measurements.
- When these issues are corrected, the values of ΔLp(2) for the RC-filter experiment fall from 1.4db down to the 0.2-0.3db range, and do not materially differ between the 3.9us and 4.7us cases. It therefore becomes extremely difficult to justify the results of the RC filter test due to nonlinear mixing.
- Paradoxially, Dr. Kunchur never mentions nonlinear mixing as a theoretical justification for his other test (speaker alignment), and in fact explicitly ascribes the results to an unknown cause. But when the corrected ΔLp(2) calculation is run, the values found are even larger than for the RC filter test, and vary quite substantially between test configurations (2.9mm: 0.76db; 6.2mm: 1.71db; 10.3mm: 3.59db). The bizarre conclusion that must be reached is that Dr. Kunchur's analyses are mutually contradictory between his two later papers (Acta Acoustica/Technical Acoustics). Either nonlinear mixing is the cause (which means the RC filter results are unexplained), or it's not (which means the RC filter's results are incorrectly justified).
- When ΔLp(2) is restated in terms of the before/after level difference of the 21khz level, and ignoring the differences in 7khz before/after levels, one finds that ΔLp(2) is strongly correlated with changes in 21khz level. If Dr. Kunchur truly wishes to justify audibility based on nonlinear mixing, and above ΔLp(2) being above a threshold, he must also admit that arbitrarily raising the 21khz level must also raise ΔLp(2) further above threshold, leading to arbitrarily lower measured temporal resolutions.
- The choice of 7khz square wave is unusually, profoundly rich in ultrasonics: The 21khz level is only 10db below fundamental, where the vast majority of tonal musical instruments have ultrasonic components far below that. That is: the input signal is entirely unrepresentative of actual music.
- Putting the above two points together yields quite possibly the most important objection to make here: Kunchur's results advocating a 5us temporal resolution are completely meaningless, because the process he uses to generate that result ought to be extended indefinitely to generate arbitrarily low resolutions. These results have no basis in psychoacoustic reality because they rely on signals further and further disconnected from realistic situations.
- Even if one discounts the nonlinear mixing justification for all of this there are still extremely good reasons to believe that the measured resolutions are inversely proportional to ultrasonic level. For instance, if one adopts a level- or peak-detection model, a signal with faster rise time ought to result in a lower measured resolution - but this faster rise time must be accomplished with increased high frequency components (keeping signal amplitude constant).
- Furthermore, his choice of 7khz square wave, far from being representative, is very carefully tuned. Any lower of a frequency would result in the 3rd harmonic becoming audible, which would invalidate his listening tests involving the manipulation of that component. Any higher of a frequency would move the 2nd harmonic further up in frequency, where human hearing is that much less sensitive; the risk becomes too great that the component would be made completely inaudible. In other words, results obtained for this type of input cannot be generalized for arbitrary music signals.
Of course, I'm only listing the comments here that relate to Dr. Kunchur's main thesis - I'm leaving out the rest of the points relating to his comments on 44.1khz digital audio, on signal synthesis, on high-end audio, etc...
ncdrawl, hold off on sending these to Kunchur just yet - I'll be able to give you/him a nicely typeset LaTeX file pretty soon with all of these points more fully fleshed out.
If I may grossly oversimplify your argument Axon, you're partially arguing that he's choosing an extreme edge case to test. However, if his intent is to map the boundaries of audibility, wouldn't an edge case be acceptable? I find the conclusion that ultrasonics are perceptible fascinating, and if he's found a case in which they actually are audible, should we not hear it out? Even though it does not represent most cases, if they are truly audible in this case, isn't that worth considering?
As an archivist, I want transparency in all cases, so I don't have to worry about the edge cases. That's why I use FLAC and not MP3. If there is any case where 44.1kHz is not sufficient, isn't that worth devising solutions for?
If I may grossly oversimplify your argument Axon, you're partially arguing that he's choosing an extreme edge case to test. However, if his intent is to map the boundaries of audibility, wouldn't an edge case be acceptable? I find the conclusion that ultrasonics are perceptible fascinating, and if he's found a case in which they actually are audible, should we not hear it out? Even though it does not represent most cases, if they are truly audible in this case, isn't that worth considering?
As an archivist, I want transparency in all cases, so I don't have to worry about the edge cases. That's why I use FLAC and not MP3. If there is any case where 44.1kHz is not sufficient, isn't that worth devising solutions for?
The use of a 7khz square wave as an input here, in this context, seems particularly unrepresentative to me, as a -10db ultrasonic third harmonic, with a signal completely absent of energy at 14khz from other sources, is a profoundly special case. It's not merely that it's an edge case - it is way, way over the edge to begin with. It's like arguing that 16 bits is insufficient because you can hear the noise with the gain raised ~20db above normal (as even shown by Meyer/Moran). Of course you can - but that situation never actually happens in the real world, where music is normalized near 0dbFS and released for an audience that actually wishes to
listen to it. More generally, Kunchur never really justifies that input signal very well, and without careful delineation, nothing's stopping anybody from boosting 21khz levels arbitrarily high to get arbitrarily low measured thresholds (like with, for instance, a bipolar pulse train).
In the final reduction ad absurdum, it's hard to tell apart his conclusions apart from a claim that (say) 200khz bandwidth is necessary for audio, because if you play extremely powerful 200khz and 202khz tones, the inevitable intermodulation is audible. The existence of
any form of intermodulation, combined with the existence of an ultrasonic bandwidth, necessarily implies that some classes of signals will show audible differences when filtered before distortion. Morevoer, this audibility will exist at
any amount of filtering greater than zero, because I'll always be able to hand you a signal that will break threshold at the intermodulation frequency.
A test with ultrasonic content at ranges more representative of real situations would restore validity, but I think that is not going to save his conclusions. In that case, if audibility is shown in the first place, it will almost certainly be above 22us. And at that point it no longer has anything to do with time resolution. But it would be a convincing proof of CD's insufficiency - but before that point is reached, the question is, how would that be possible when every prior attempt has failed?
As an archivist, I want transparency in all cases, so I don't have to worry about the edge cases. That's why I use FLAC and not MP3. If there is any case where 44.1kHz is not sufficient, isn't that worth devising solutions for?
I think his results are really interesting, but to an archivist, they're not relevant until they're shown to apply to something approaching actual audio. After all if you just want to store square waves, you shouldn't be using PCM in the first place because of its nasty requirement that signals be band limited . . .
I think his results are really interesting, but to an archivist, they're not relevant until they're shown to apply to something approaching actual audio. After all if you just want to store square waves, you shouldn't be using PCM in the first place because of its nasty requirement that signals be band limited . . .
…Because other
sampling methods don't require band-limited signals?
For archival purposes there is a decent argument for going beyond redbook "just in case" ... Perhaps the lizard people who will take over the earth after we nuke ourselves will have decent ultra-sonic hearing and want the full experience? It's not like the behaviours of professional archivists have much relationship to the behaviour of normal people anywhere else. (Or will you be micro-scribing my message onto a nickel plate?)
I think his results are really interesting, but to an archivist, they're not relevant until they're shown to apply to something approaching actual audio. After all if you just want to store square waves, you shouldn't be using PCM in the first place because of its nasty requirement that signals be band limited . . .
…Because other sampling methods don't require band-limited signals?
Sampling a square wave is silly. Just store the frequency, duty cycle, phase and amplitude. 4 numbers and you don't have to band limit it!
I'd like to add a few comments regarding the "Temporal resolution of hearing probed by bandwidth restriction" paper.
One potential issue is the second-harmonic contribution of the test setup. Since this contribution is within the audible band (14 kHz), special attention needs to be paid to it. Of course, an ideal square wave has no even-order harmonics by virtue of its half-wave symmetry, i.e. f(t +/- T/2) = -f(t). But there are ways that the second harmonic can creep back in. Two ways I can think of are the duty cycle of the square wave not being exactly 50 percent, and second-harmonic distortion of the test setup's transducers and electronics. Either of these situations will eliminate the half-wave symmetry and introduce even-order harmonics. The text below figure 1 states "The acoustic output from the transducer is devoid of even numbered harmonics because of the square-wave signal fed to it". Of course, in the real world, that signal can't be entirely devoid of second harmonic, and indeed figure 4 shows its presence. Oddly, figure 4 shows the frequency components in terms of power, where dB would have been better if clarity were the intent. In any case, the power ratio of second harmonic to fundamental is shown as 1e-6, giving a voltage ratio of 1e-3 (-60 dB) at the mic preamp output where this measurement was presumably taken.
One interesting thing that can be done is to make the generous assumption that the electronics and transducers have zero second-harmonic distortion, and assume this second-harmonic component is due entirely to the duty cycle of the square wave not being exactly 50 percent. One could then derive the Fourier series coefficients of a rectangular wave having duty cycle d, where 0 < d < 1. Then one could figure out what values of d correspond to a second-harmonic component 60 dB down from the fundamental. This would correspond to a "best case" scenario, because with electronics and transducers having non-zero second-harmonic distortion, the tolerance on the duty cycle of the square wave would have to be even tighter to take into account those additional second-harmonic components. Suppose we have a rectangular wave with a symmetrical voltage swing of +/- Vp and a duty cycle d. It's not too hard to show that the Fourier series coefficients vn of this function are:
vn = (4Vp/(pi*n)) sin(pi*d*n) for n >= 1 (i.e. this excludes the DC term)
The ratio r of the second harmonic to the fundamental is:
r = (1/2) sin(2*pi*d)/sin(pi*d)
Now we can solve numerically for the value of d that makes r = +/- 1e-3. This will give two answers - one slightly less than 0.5 and one slightly greater. Plugging this into MathCad, we get r1 = 0.49968 and r2=0.50032. So the allowable range of the duty cycle is 50 +/- 0.032 percent. This is an extremely stringent requirement, and yet this number is optimistic for two reasons. The first is the previously mentioned assumption that the electronics and transducers have no second-harmonic distortion. The second is that real-world square wave generators use nonlinear circuits which in general will have slightly asymmetric rise and fall characteristics. The Fourier series coefficients assume zero rise and fall times. I'm finding it hard to believe that he actually achieved this number, especially with two transducers involved (the headphones and microphone). So let's look at table 1, where he lists the harmonic components of the transducer output to see what he actually measured for the second harmonic in each condition. Well, those measurements aren't there, only the fundamental and the third and fifth harmonics. The entire issue is papered over with the statement "The acoustic output from the transducer is devoid of even numbered harmonics because of the square-wave signal fed to it". Since changes in this second harmonic value could explain the experimental results without needing some hypothesis regarding the alleged ability of the ear to detect signals above the frequency limit of human hearing, it's essential to include these data. Yet he fails to do so.
So, on that note.... an off topic comment. I just realized that every chiptunes/NSF/etc player probably has a broken square wave generator.
I checked the source code for one (nosefart) and it just does the naive synthesis. I checked blapp's libraries, which are used in many places, including foo_input_gep, and go to some trouble to ensure bandlimited synthesis... If I make an NSF with nothing but an A#7 square wave, and decode it through that library, it only rejects the 26khz harmonic to the tune of -50db!
It's kind of comical that the "high quality" players support oversampling... like, say, to 96khz. Whereas to get the aliasing under the 16-bit noise floor, one would have to sample at 1.45Ghz.
That's just barely infuriating enough for me to consider hacking together a player that relied entirely on additive synthesis.
It may well be that Dr. Kunchur's audio work is getting more thorough 'peer review' here than he's gotten so far. ;>
It shall be interesting to see the response.
I'd like to add a few comments regarding the "Temporal resolution of hearing probed by bandwidth restriction" paper.
One potential issue is the second-harmonic contribution of the test setup. Since this contribution is within the audible band (14 kHz), special attention needs to be paid to it. Of course, an ideal square wave has no even-order harmonics by virtue of its half-wave symmetry, i.e. f(t +/- T/2) = -f(t). But there are ways that the second harmonic can creep back in. Two ways I can think of are the duty cycle of the square wave not being exactly 50 percent, and second-harmonic distortion of the test setup's transducers and electronics. Either of these situations will eliminate the half-wave symmetry and introduce even-order harmonics. The text below figure 1 states "The acoustic output from the transducer is devoid of even numbered harmonics because of the square-wave signal fed to it". Of course, in the real world, that signal can't be entirely devoid of second harmonic, and indeed figure 4 shows its presence.
Well, given that he actually measured the acoustic field at the listening position and no even harmonics were observed at that point, I figure that is reasonable enough evidence that they do not exist... but read on.
Oddly, figure 4 shows the frequency components in terms of power, where dB would have been better if clarity were the intent. In any case, the power ratio of second harmonic to fundamental is shown as 1e-6, giving a voltage ratio of 1e-3 (-60 dB) at the mic preamp output where this measurement was presumably taken.
In his first paper (to Proceedings of Meetings on Acoustics) Kunchur explicitly rejects the use of FFTs in the power spectrum analysis:
An enormous time (of the order of two years) and effort were spent to develop the instrumentation and the methods for checking for artifacts. For example, for just the Fourier spectrum shown in Fig. 4, it took a few months to develop the instrumentation setup and to write the C code (FFT was not used). To measure one such spectrum takes over a week.
IIRC, when questioned about this he said the implementation was an "elementary" one derived from textbook sources. I assert that anything using a power spectrum that can't use an FFT is unquestionably
not elementary.
Well, given that he actually measured the acoustic field at the listening position and no even harmonics were observed at that point, I figure that is reasonable enough evidence that they do not exist... but read on.
Figure 4 is evidence that they
do though. This is not a surprise in any way. Ever try to get clean square waves to a scope from a pulse generator through a cable as short as 1m? A coworker of mine years ago thought his pulse generator was defective because of all the ringing and other artifacts observed on a scope when doing this. He was going straight into the high-impedance input of the scope. What he didn't know was that the cable had to be terminated in its characteristic impedance at the scope end. Once that's done, a "textbook" square wave is obtained. Otherwise, it's an awful-looking mess.
When the theory shows the second harmonic component to be so strongly dependent on tiny changes in the duty cycle, it raises concerns as to whether that was really controlled adequately. I don't see evidence of any such precautions having been taken.
So, on that note.... an off topic comment. I just realized that every chiptunes/NSF/etc player probably has a broken square wave generator.
Sweet! I get to pester kode54! We'll see how far I can go with this.
Where did you come up with that 1.45GHz figure? The 5A02 was clocked at closer to 1.6MHz, and thus produced sound at about that rate.
Well, I'm thinking more specifically of the theoretical case, where the harmonics decay at An=1/n, so to get -96db down....
…Because other sampling methods don't require band-limited signals?
Oh, what sampling methods are we talking about here?
I may be incorrect, but assuming a signal model different from the traditional (sum of sines), isn't it possible to come up with better sampling and reconstruction functions than the (time-domain) sinc? If yes, each time-domain sample could potentially be multiplied with a non-band-limited function (as opposed to the sinc) and thus produce a non-band-limited reconstructed signal.
In the square-wave case, certainly one could come up with better reconstruction rules given a signal model where square-waves have a high probability, mitigating e.g Gibbs issues.
Btw. does anyone know of references to more genereal work in this field? How to find the optimal sampling and reconstruction functions given a signal model that is less trivial than just "band-limited"?
/Pontus
…Because other sampling methods don't require band-limited signals?
Oh, what sampling methods are we talking about here?
I may be incorrect, but assuming a signal model different from the traditional (sum of sines), isn't it possible to come up with better sampling and reconstruction functions than the (time-domain) sinc?
You are correct. PCM is general, but typically not optimal if you have extensive prior information about a signal. In this case, as I hinted before since you know theres only one type of signal you (perfect square waves) you can perform exact non-bandlimited sampling while using a finite (and absolutely tiny) number of measurements.
In the square-wave case, certainly one could come up with better reconstruction rules given a signal model where square-waves have a high probability, mitigating e.g Gibbs issues.
Correct again. Technically PCM doesn't even work for square waves since they're not band limited.
Btw. does anyone know of references to more genereal work in this field? How to find the optimal sampling and reconstruction functions given a signal model that is less trivial than just "band-limited"?
"Compressed sensing", "compressed sampling" and "sparse sampling" are wildly used in engineering and optics to refer to other modes of sampling apart from Nyquist-Shannon. If you google you can find texts covering various approaches.
Btw. does anyone know of references to more genereal work in this field? How to find the optimal sampling and reconstruction functions given a signal model that is less trivial than just "band-limited"?
"Compressed sensing", "compressed sampling" and "sparse sampling" are wildly used in engineering and optics to refer to other modes of sampling apart from Nyquist-Shannon. If you google you can find texts covering various approaches.
Thanks for your answers and the pointers. I've studied those concepts briefly and they seem to consider mostly the case of modifying both the sampling and reconstruction functions. However I haven't found any literature describing how to modify the reconstruction function, given standard (uniform Dirac pulses) sampling but a signal model other than "band-limited".
It may well be that Dr. Kunchur's audio work is getting more thorough 'peer review' here than he's gotten so far. ;>
It shall be interesting to see the response.
Why you insist on spewing such counterproductive nonsense is beyond me. I have spoken with Dr. Kunchur personally (via phone) and know that his work has been rigorously reviewed by authorities in many fields. He describes this process/experience in great detail in his writings.
thanks.
OK, OK... But I must say I find nothing rude in Krabapples comment. In fact (without prior knowledge of the case at hand) it seems plausible.
I'm not implying anything w.r.t. Dr. Kunchur's papers here, but in my experience peer review process can mean very different things in practice. Factors such as journal ranking, mood, interest and workload of reviewer in combination with reputation of author (or co-authors), novelty of research field etc. can influence the process. After all, reviewers are human beings and can have a bad day.
Sometimes you get a very thorough review with lots of constructive comments, sometimes the reviewer just seems to have made a few read-throughs with minor remarks and that's it. Sometimes you may even find errors in your own output which passed unnoticed through the peer-review process. From my graduate student years (semiconductor laser physics) I remember seeing a few papers every now and then in rather prestigious journals (e.g. Appl. Phys. Lett.) that I would never have let through myself.
So, the fresh eyes provided by Axon and others here at HA may see things from another angle and discover things unnoticed by the peer reviewers. It can only be a good thing.
Could you please stop being confrontational and rude??
He might have been so elsewhere, but there's really little to critique about his posts in this thread. Your response is more "confrontational and rude" than his initial post. He was just commenting that he expects to see some engaging peer-review here, which is a very valid belief.
I am a member of the sphile forum, and do not appreciate his attacks.
................and we shall know them by the fruits that they bear.
I think it's reasonable to question the peer review of some of the quotes that you guys over there have mined, because they are just wrong.
It may not have been what the Dr. intended to mean, but nipping that kind of misapprehension is a substantial part of what reviewing and editing is on about.
The quotes wherein the Dr. is alleged to have said that amplitude resolution has no bearing on time resolution in a PCM signal is also cause for questioning. Now, I wasn't there, I don't know exactly what the good Dr. said, but if he said that, he's sorely mistaken.
I read his reply. It is equivocation, plain and simple. The words have defined, specific meanings in the field, and part of peer review is making sure that the author uses words in such a fashion. So, when that does not happen, that's yet another reason to be concerned about the review process and the reported results.
um...no. you are wrong.. he was implying that the peer review Dr. Kunchur has gotten thus far was bogus (as indicated by his quotes around the words "peer review")...
There is a distinct possibility it was reviewed by the wrong set of "peers', I must say.
Hence the uproar on Stereophile's forum, where it's delightful to see what fulsome respect the letters 'PhD' can garner from audiophiles when they really want to believe
I am a member of the sphile forum, and do not appreciate his attacks.
Bismarck, or someone, said that you'd not feel the same about laws or sausages if you'd seen them being made. Add to that list PhDs. Some PhDs, of course, are awarded for outstandingly original work; most, for good solid stuff which shows that the candidate has learned how to do research (in the specified field); some are awarded for no visible reason. Feynman once said that he didn't supervise PhDs because the degree was awarded for work by a member of faculty completed under particularly trying circumstances. Harsh, but everyone who's supervised PhDs laughs when they hear it. Rule of thumb: anyone who makes a big deal of their doctorate, for any reason other than trying to get upgraded on an airline, is to be viewed with suspicion.
It may well be that Dr. Kunchur's audio work is getting more thorough 'peer review' here than he's gotten so far. ;>
It shall be interesting to see the response.
Why you insist on spewing such counterproductive nonsense is beyond me. I have spoken with Dr. Kunchur personally (via phone) and know that his work has been rigorously reviewed by authorities in many fields. He describes this process/experience in great detail in his writings.
so....Could you please stop being confrontational and rude?? I mean...firs the jab at the Stereophile forum, now another jab at Dr. Kunchur.
If you aren't going to be polite(like the rest have), please do not pollute this thread as you have elsewhere.
your comments have added nothing.
thanks.
Based on talking to Dr. Kunchur, *you* are convinced his work has been thoroughly vetted by relevant authorities.
Your confidence might matter to me if you can answer 'yes' to all of the following questions:
Are you demonstrable expert in the fields related to the audio work Dr. Kunchur has published?
And if not,
Are you a scientist? Are you familiar with peer review -- its potential and demonstrated strengths and flaws? Are you equipped to even begin to assess the bona fides presented for a scientist's work, e.g. his CV, the list of conferences he's presented it at, the journals he's published it in? Would you even know how to gauge their relevance to a particular bit of his published research? How to determine if he publishes in 'good' journals or not? How to find out what esteem his work is held in by people working in the same field?
If you can't answer that last set in the affirmative, what is *your* confidence supposed to mean to me, someone who *can* answer yes to all of those?
I am a scientist . Not famous, not especially accomplished. My field is not digital audio signal processing or psychoacoustics. But I have certain skills to investigate claims from a field outside my own. These aren't unusual skills for scientists; you pick them up of necessity. And looking into where he's presented , and where he's published, I'm not 100% convinced Dr. Kunchur's audio work has gotten rigorous peer review from properly targeted reviewers.
I might add my own proverb that when a scientist publishes well outside his or her main field of expertise, as Dr. Kunchur has done with his audio work, he or she stands a good chance of either finding something fresh...or making a fundamental error. Some people here who are pretty well versed in matters closely related to his work, appear to think he's done the latter. That gives me pause.
If you find that 'impolite' maybe you need to attend a few scientific conferences yourself, or try being part of a peer review process. It's not for the faint of heart or for those with a romantic idea of how science gets done ('how the sausage gets made')
But even if you disagree, it should be clear even to you that there is vastly more detailed and substantive review of Dr. Kunchur's audio work being offered here at HA that on offer from certain braying asses on Stereophile's forums who merely keep repeating HOW DARE YOU!! LOOK AT THE LIST OF CONFERENCES!!! LOOK AT THE PHD!!! LOOK HE HAS TWO PAPERS!! SO SHUT UP, DR KUNCHUR RULEZ!!!!
So as I said: I look forward to Dr. Kunchur's responses and debate on the technical critiques offered here...assuming he reverses his decision that all substantive issues have already been addressed.
blah snip blah
what are your credentials?
One need not be a scientist to recognize rudeness. You are acting like a jerk. I could be an illiterate fool and still be spot on about your being antagonistic.
it was pointed out over there, and now I am pointing it out here.
this place is a canoe club of sorts so you may get by with it more... but ill call a spade when I see it...Herr Spade.
Yikes.
I'm an employed scientist (honestly, I'd be tempted to say we're 'technical professionals' on the production side, since original research isn't the primary thrust) as well, and looking at the questions posed on some of the other forums, and the response from Dr. Kuncher, I'd say he has a good point or three even assuming badly forged credentials scrawled with blunt crayons on toilet tissue.
What are
your credentials and how is
your core specialty relevant to psychoacoustics and signal processing?
To be perfectly honest, my credentialed (http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=imm.TOC&depth=2) competencies (http://en.wikipedia.org/wiki/Chemistry) don't relate directly to these topics, but I do have an amateur interest and the circumstances here are interesting. The researcher of the (possibly revelatory, if documented accurately and actually reproducible) work in question here has called an end to participation in even cursory internet discussions (with numerous apparently relevant unanswered questions, some posed by figures known in the specific field) and taken his ball and gone home to his core competency. There is no doubt in my mind that talented superconductivity researchers are a much needed force in society (buddha knows we could use some nice 250mph maglev trains here on the west coast), but phonon dynamics (http://en.wikipedia.org/wiki/Phonon) have at best a tangential relevance to this topic. Given the sweeping conclusions one might be tempted to draw from Kuncher's statements regarding this research, I'm a bit befuddled as to his reticence to engage in discussion.
Therefore, with all due respect, pimping Dr. Kuncher's general scholastic record is less interesting and
vastly less relevant than actually addressing the specific technical questions posed.
There is also an apparent 'context' here, which colors both Kuncher's work and is reminiscent of that surrounding some of his
references (re: ultrasonics).
edit: formatting, redundancy
I am not upset. I was simply pointing out the guy's personality flaws.
Let's keep the ad hominem to a minimum. Behavior like this from either side is unacceptable, especially in the "Scientific Discussion" forum.
This thread is for technical discussion of the paper put forward by Dr. Kunchur. I may not personally have a doctorate in a related field, but that is a goal of mine. This discussion is basically a form of peer review. There are quite a few people here that have significant degrees of understanding in this field, ncdrawl. Many professionals, scientists, passionate amateurs, and everywhere in-between.
Why not just just create a 7.35kHz square wave "manually", i.e. for 44.1kHz sample rate 6x 32767 followed by 6x -32768 <repeat>; 15x for 110.25kHz and 24x for 176.4kHz?
It still aliases, just that the aliased harmonics overlap onto other valid harmonics, so the distortion doesn't sound as bad. You can get better square wave approximation through other means like minimum-phase bandlimited steps (minBLEP), band-limited impulse trains (BLIT), or windowed sinc method (BLIT-SWS). But only exact synthesis (sum of sines) is going to be exact.
Why not just just create a 7.35kHz square wave "manually", i.e. for 44.1kHz sample rate 6x 32767 followed by 6x -32768 <repeat>; 15x for 110.25kHz and 24x for 176.4kHz?
It still aliases, just that the aliased harmonics overlap onto other valid harmonics, so the distortion doesn't sound as bad.
In this case, either the fundamental and harmonics end up at the amplitude they should be, or they don't.
Are you saying that they
don't?
Of course you end up with something where the phase relationship between the square wave and the sampling instants is specific and fixed - but can't you get exactly the same thing by summing sines? (or nearly the same thing, assuming the comparison is with this "square" wave run through a decent but real world oversampling DAC).
Cheers,
David.
Why not just just create a 7.35kHz square wave "manually", i.e. for 44.1kHz sample rate 6x 32767 followed by 6x -32768 <repeat>; 15x for 110.25kHz and 24x for 176.4kHz?
It still aliases, just that the aliased harmonics overlap onto other valid harmonics, so the distortion doesn't sound as bad.
In this case, either the fundamental and harmonics end up at the amplitude they should be, or they don't.
Are you saying that they don't?
Of course you end up with something where the phase relationship between the square wave and the sampling instants is specific and fixed - but can't you get exactly the same thing by summing sines? (or nearly the same thing, assuming the comparison is with this "square" wave run through a decent but real world oversampling DAC).
Cheers,
David.
Sorry for the confusion. I meant that, if the sampling rate is a multiple of the square wave's fundamental frequency, then the aliased harmonics will have the exact frequency as a non-aliased harmonic.
That is, if a square wave has harmonics at f1 through f<infinite>, nyquist frequency (fn) will also be a harmonic, and every harmonic fn+1 will alias to fn-1 which is also a harmonic.
It's still aliasing distortion, but it "sounds" OK because it is still harmonic. Just that the spectral distribution of the waveform will be different than a proper sampled (or properly synthesized) version.
A real unaliased square wave, when digitized, will have gibbs effect during the transitions at all frequencies (and the length of ringing is frequency-invariant), because the impulse in the analog domain turns into a sinc wave in the digital domain.
And, to clarify. Yes, a "naive" digital square wave will have the wrong amplitude for its harmonics even when the sampling rate is a multiple of the fundamental frequency.
Yes, I see it now (having tried it!), it's looks (after the DAC) and sounds almost the same, but there's too much of the high frequency harmonics because the aliases are added in.
Cheers,
David.
I am not upset. I was simply pointing out the guy's personality flaws.
Let's keep the ad hominem to a minimum. Behavior like this from either side is unacceptable, especially in the "Scientific Discussion" forum.
I have no problem complying with that, assuming that the SOP is enforced in an even-handed manner.
I love seeing healthy discussion though, my entire reason for posting the papers..
From the first paragraph at: www.physics.sc.edu/kunchur/Acoustics-papers.htm (http://www.physics.sc.edu/kunchur/Acoustics-papers.htm)
Our recent behavioral studies on human subjects proved that humans can discern timing alterations on a 5 microsecond time scale, indicating that that digital sampling rates used in common consumer audio (such as CD) are insufficient for fully preserving transparency.
This statement contains a glaring error:
44.1 kHz sampling PCM systems are perfectly capable of reproducing the phase of audible frequencies to picosecond accuracy. The need for 5-microsecond temporal accuracy does NOT indicate the need for a higher sample rate. It simply indicates that jitter must be less than 5-microseconds.
From the summary paragraph in "Audibility of temporal smearing and time misalignment of acoustic signals" (http://www.physics.sc.edu/kunchur/papers/Audibility-of-time-misalignment-of-acoustic-signals---Kunchur.pdf) by Dr. Kunchur
These qualitative and anecdotal observations point to the possibility that human hearing may be sensitive to temporal errors, τ , that are shorter than the reciprocal of the limiting angular frequency [2πfmax]^−1 ≈ 9 μs, thus necessitating bandwidths in audio equipment that are much higher than fmax in order to preserve fidelity.
Again, the same glaring error:
The Nyquist frequency does NOT determine the temporal resolution of a PCM system. The entire premise of the paper is fundamentally flawed. The experiment proves the need for moderately controlled jitter but does not establish the need for higher sample rates.
I have no problem complying with that, assuming that the SOP is enforced in an even-handed manner.
If you notice us not being even-handed, you are welcome to PM me personally, though there are certain technical points that are assumed to be true here on this forum. Asserting them usually does not invoke Term of Service #8, though asserting the contrary will. Nevertheless, you're welcome to assert the contrary if you have back-up and can make a solid case.
From the summary paragraph in "Audibility of temporal smearing and time misalignment of acoustic signals" (http://www.physics.sc.edu/kunchur/papers/Audibility-of-time-misalignment-of-acoustic-signals---Kunchur.pdf) by Dr. Kunchur
These qualitative and anecdotal observations point to the possibility that human hearing may be sensitive to temporal errors, ? , that are shorter than the reciprocal of the limiting angular frequency [2?fmax]^?1 ? 9 ?s, thus necessitating bandwidths in audio equipment that are much higher than fmax in order to preserve fidelity.
Again, the same glaring error:
The Nyquist frequency does NOT determine the temporal resolution of a PCM system. The entire premise of the paper is fundamentally flawed. The experiment proves the need for moderately controlled jitter but does not establish the need for higher sample rates.
That seems too harsh to me - I agree that Dr. Kunchur is making huge conceptual mistakes with his treatment of digital audio, and in a lot of other things for that matter, but the experiment itself can't be dismissed so easily. Despite all the issues I have with the paper that I mentioned earlier, isn't it at least plausible on some sort of a priori level that, if one can guarantee a distortion of fairly reasonable consistency (internal to the ear), that testing of ultrasonics on that basis
may be reasonable?
This statement contains a glaring error:
Welcome to the hornets' nest .
In all seriousness, he does attempt to address this in his FAQ document, in the last sentence of page 1 and the first couple of sentences of page 2. He states:
If there are two sharp peaks in sound pressure separated by 5 microseconds (which was the threshold upper bound determined in our experiments), they will merge together and the essential feature (the presence of two distinct peaks rather than one blurry blob) is destroyed. There is no ambiguity about this and no number of vertical bits or DSP can fix this. Hence the temporal resolution of the CD is inadequate for delivering the essence of the acoustic signal (2 distinct peaks).
Of course, you might wonder what the relationship of his test (audibility of a 5us time constant filtering a 7 kHz square wave) is to the "two distinct pulses separated by 5us" scenario. If so, you are not alone. However it does seem to bear a closer relationship to his experiment with the two loudspeaker drivers where the delay between them was varied to find the minimum detectable delay difference. But this has been criticized on the grounds that the spectrum of the close-together pulses vs. that of the pulses further separated can differ by an audible amount below 20 kHz.
That seems too harsh to me - I agree that Dr. Kunchur is making huge conceptual mistakes with his treatment of digital audio, and in a lot of other things for that matter, but the experiment itself can't be dismissed so easily. Despite all the issues I have with the paper that I mentioned earlier, isn't it at least plausible on some sort of a priori level that, if one can guarantee a distortion of fairly reasonable consistency (internal to the ear), that testing of ultrasonics on that basis may be reasonable?
I do not dismiss the test results that attempt to quantify our ability to hear abrupt temporal shifts. This is useful information. However, the paper assumes that these 5 usec temporal shifts cannot be represented in a 44.1 kHz PCM system. This assumption is simply false.
Dr. Kunchur is... and in a lot of other things for that matter, but the experiment itself can't be dismissed so easily. ... isn't it at least plausible on some sort of a priori level that, if one can guarantee a distortion of fairly reasonable consistency (internal to the ear), that testing of ultrasonics on that basis may be reasonable?
I agree totally. The jumping to conclusions re digital audio is flawed, but that does not invalidate
both experiments, and there might be something interesting lurking there.
By the way:
- When these issues are corrected, the values of ?Lp(2) for the RC-filter experiment fall from 1.4db down to the 0.2-0.3db range, and do not materially differ between the 3.9us and 4.7us cases. It therefore becomes extremely difficult to justify the results of the RC filter test due to nonlinear mixing.
Isn't that already in the RC paper, implicitly? One of the test subjects was tested to have hearing capability
to only less than 10kHz, while he faultlessly detected RCs down to 5.6us, and scored 8/10 on 4us (or whatever the figures were).
That seems too harsh to me - I agree that Dr. Kunchur is making huge conceptual mistakes with his treatment of digital audio, and in a lot of other things for that matter, but the experiment itself can't be dismissed so easily. Despite all the issues I have with the paper that I mentioned earlier, isn't it at least plausible on some sort of a priori level that, if one can guarantee a distortion of fairly reasonable consistency (internal to the ear), that testing of ultrasonics on that basis may be reasonable?
Well, that's the whole point - but this flawed experiment didn't probe this properly - and all previous proper experiments show that, apart from via bone conduction, ultrasonics don't distort to create audible frequencies
in the ear.
In equipment, yes, but not in the ear.
If someone proves otherwise (quite possible), it'll be interesting.
IIRC there was someone here who did (playing ultrasonic from a separate audio system(!) and ABXing presence / absence in the presence of an audible sound), but I didn't see how the thread ended.
Cheers,
David.
I'm just coming back from vacation. Thanks for the many insightful posts, I have just learned a lot! Until now I always thought that PCM being bandwidth limited and requiring low pass filtering is just about frequencies that no human ear would care about anyway and no big deal else.
Comprehending that the plot of a properly encoded PCM bitstream of a square wave is not square, why low passing before ADC is so important, and why perfect square waves don't fit into PCM in theory took me a big leap forward.
I had already checked DVD-A prices for my favorite records online... But until I see Kunchur refuting some of this thread's objections I'll stay as happy with Redbook as I always have been.
But until I see Kunchur refuting some of this thread's objections I'll stay as happy with Redbook as I always have been.
As you seem to be saying, there's a lot of interesting twists along the way, but getting back to basics is very often a good thing.
Setting up a proper listening test that shows that high sample rates do make a difference with music is simpler and less costly than ever.
There is now a ton of high sample rate audio kicking around for free or not much money. Actually setting up the experiment can be accomplished with not only reasonably-priced buyware, but even with freeware.
There seems to be only one minor thing lacking - a pair of ears that produce results that are positive for high sample rates.
Whenever we see a paper like Kunchur's, we need to get back to basics. The basics are that just about anybody who wants to can set up a reliable listening test involving readily-available musical program material that has the potential to produce a positive outcome. Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Dr. Kunchur will be posting confutations in form of FAQ updates...will be doing that regularly.
Apparently he has been bestowed with a great deal of common sense to go with his impressive educational credentials... IE he avoids online arguments. Very smart decision.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Dr. Kunchur will be posting confutations in form of FAQ updates...will be doing that regularly.
So, when's the next regular installment?
Apparently he has been bestowed with a great deal of common sense to go with his impressive educational credentials...
IE he avoids online arguments. Very smart decision.
Possibly, but perhaps not for the reason you think. The closer to 'real time' an interrogation by knowledgeable critics is, the more stressful it can be. In a FAQ one is in total control of which questions will be addressed and how in-depth the address will be. On a forum you don't have that control.
Dr. Kunchur will be posting confutations in form of FAQ updates...will be doing that regularly.
Might we ask then that he keeps/publishes a record of the changes, so that
we don't have to re-read all all of the time?
Further, there already is something interesting in an earlier
version of the FAQ document that is missing in the present one, if I remember
correctly.
It had to do with the time accuracy of 44.1kHz sampled systems ;-)
> Dr. Kunchur will be posting confutations in form of FAQ updates...will be
> doing that regularly.
That will be interesting to see if he keeps digging himself deeper into a hole or starts trying to climb out. Unless he is a completely lost in audiophile belief (he is clearly a believer) he must have twigged by now that his lack of understanding of the information contained in a set of samples has lead him to make a bit of a fool of himself among his peers. But he is getting attention from the audiophile community. So which is more important to him? We will see.
> Apparently he has been bestowed with a great deal of common sense to go with
> his impressive educational credentials...
Why do you find them impressive? If you find scientific credentials impressive, why do you discard the bulk of relevant scientific work on sound perception performed by people with a higher standing in the field than your good Dr.?
> he avoids online arguments. Very smart decision.
In his position it would be rather unwise to do otherwise.
Regarding my earlier post (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=73598&view=findpost&p=648796) about the lack of second harmonic data in the "Temporal resolution of hearing probed by bandwidth restriction" article, it looks like he published those data in another article, "Probing the temporal resolution and bandwidth of human hearing" (PDF file here) (http://www.physics.sc.edu/kunchur/papers/Probing-the-temporal-resolution-and-bandwidth-of-human-hearing--Kunchur.pdf). I had previously used the number 1e-3 for the second harmonic relative to the fundamental, read from his graph in figure 4 of "Temporal resolution of hearing probed by bandwidth restriction", to compute the required accuracy for the duty cycle of the square wave. His actual tabular data show a second harmonic relative to the fundamental of 0.0003, or -70.4 dB, rather than -60 dB as I had originally read from his graph. Plugging in the numbers, the previous requirement for the duty cycle accuracy, which was 50 +/- 0.032 percent, now becomes 50 +/- 0.0095 percent for a -70.4 dB second harmonic. That assumes zero second-order distortion in the buffer that follows the square wave generator, the buffer after the LPF, the headphone amplifier, the headphones themselves, the microphone, the mic preamp, and the 12-bit A/D converter in the oscilloscope used to digitize the data. I've worked with test equipment for many years, and in all that time I've never seen a square wave generator specified for its second harmonic content. The specifications for the pulse generator he used (http://www.globalspecialties.com/4001_spec.html) certainly don't show anything like that. I suppose it would be possible to put it in the variable pulse width mode and adjust its parameters to minimize the second harmonic, but according to Axon's communications with Dr. Kunchur, the spectrum computation takes days(?). That would make such an approach impractical. Or it could actually be that the square wave symmetry of the generator is that good. It seems unlikely though.
On the plus side, as long as this ratio were low enough, and constant with varying bandwidth, it probably wouldn't influence the results at all. The very low values of second harmonic just jumped out at me, considering how tiny errors in duty cycle of the square wave (and other contributors to errors in half-wave symmetry) result in non-negligible second harmonic spectral content, as well as how the actual data for the second harmonic depend on unspecified parameters of the square wave generator. Also, the second harmonic data in figure 4 of "Temporal resolution of hearing probed by bandwidth restriction" looks like a noise blip, rather than having a wide baseline as the other harmonics do - which also looks odd. I suppose this could be due to the resolution limit of the 12-bit converter used in the scope.
Dr. Kunchur will be posting confutations in form of FAQ updates...will be doing that regularly.
Might we ask then that he keeps/publishes a record of the changes, so that
we don't have to re-read all all of the time?
Further, there already is something interesting in an earlier
version of the FAQ document that is missing in the present one, if I remember
correctly.
It had to do with the time accuracy of 44.1kHz sampled systems ;-)
Oh, man. I completely forgot about that. I've noticed the same thing. I have a printout of his original FAQ, so at least we'll know exactly what he's changing.
BTW, I find this behavior highly unprofessional. I would almost find it tantamount to lying.
BTW, I find this behavior highly unprofessional. I would almost find it tantamount to lying.
Little sympathy for a man with problems? His beliefs are not holding up, the basis for his audiophile fame is starting to crumble, he has published papers in his name with erroneous content, how to fix the obvious false statements in his FAQ but still give audiophiles what they want to see?, should he attempt to keep going? read a bit about sampling and then correct the mistakes? delete the FAQ, remove the links to the hardcopies of the papers and keep quiet for a bit,...
It is interesting to see what he will do and if he ends up trusting existing scientific knowledge a bit more and audiophile beliefs a bit less.
I need to do further reading to do a full refutation, but I think I figured out where the problem is right.
Dr Kunchur shows that humans can discern temporal resolution down to 5 microseconds. I mostly agree with this - it's a fundamental part of pitch perception. Where he makes the mistake is assuming that 44,100Hz sampling rate only has a temporal resolution of 22 microseconds. This is patently false. 44100Hz audio has far more temporal resolution than this, due to the antialiasing filter. If you model the impulse response function of the antialiasing function as a sinc wave, small (<20microseconds) delays in an impulse event will correspond to vastly different outputs from the sinc() function. If the input impulse is perfectly aligned, the sinc() output will give 1 followed by an infinite amount of zeroes, but a half-sample (11 microsecond) delay will give 2/PI followed by a series of decaying values. The reconstruction filter will output the original half-sample delayed signal, but in bandlimited form. The half-second delay on the impulse will still be present on the output signal.
Also, for his FAQ example of two peaks separated by 5 microseconds, my statement still holds true. The two peaks will "ring" the anti-alias filter in such a way that those two peaks still have an impact into the digital sampling (and on the eventual post-reconstruction-filter output). The waveform will look different and the high frequency spectra will be lost but the "information" of the two separate peaks will remain in the final analog output.
two peaks separated by 5 microseconds, my statement still holds true. ...The waveform will look different and the high frequency spectra will be lost but the "information" of the two separate peaks will remain in the final analog output.
Er, no. They would be merged into one peak. Do the experiment.
But this of course does not prove that 44.1kHz is insufficient because of the simple
fact that it has not been proven that an audio system can only be audibly transparent
when it keeps 5us impulses separated. That's the error.
It's a bit like stating that airplanes are rubbish because you can't do space missions
with them.
two peaks separated by 5 microseconds, my statement still holds true. ...The waveform will look different and the high frequency spectra will be lost but the "information" of the two separate peaks will remain in the final analog output.
Er, no. They would be merged into one peak. Do the experiment.
They're still separate in the frequency domain.
two peaks separated by 5 microseconds, my statement still holds true. ...The waveform will look different and the high frequency spectra will be lost but the "information" of the two separate peaks will remain in the final analog output.
Er, no. They would be merged into one peak. Do the experiment.
But this of course does not prove that 44.1kHz is insufficient because of the simple
fact that it has not been proven that an audio system can only be audibly transparent
when it keeps 5us impulses separated. That's the error.
It's a bit like stating that airplanes are rubbish because you can't do space missions
with them.
Please re-read the part I just bolded in the "quote" block above.
But this of course does not prove that 44.1kHz is insufficient because of the simple
fact that it has not been proven that an audio system can only be audibly transparent
when it keeps 5us impulses separated. That's the error.
What do these two peaks in the time domain represent? Without a formulation of a comprehensible hypothesis how can one perform a test and claim an error? (Ditto for benski.)
But this of course does not prove that 44.1kHz is insufficient because of the simple
fact that it has not been proven that an audio system can only be audibly transparent
when it keeps 5us impulses separated. That's the error.
What do these two peaks in the time domain represent? Without a formulation of a comprehensible hypothesis how can one perform a test and claim an error? (Ditto for benski.)
They are supposed to repesentation two sounds occuring nearly, but not quite, simultaneously, for example two drums where the mallets hit 5 microseconds apart.
The proper way to test would be digitize the sound @ 44.1khz and again at something much higher like 384khz (which has a sampling period smaller than 5 microseconds), and do a blind ABX test with a set of listeners (using the same criteria for choosing listeners as the experiment in question). Most of us at HA know better than to trust a spectrogram. Although my point was simply that the post-experiment statement of the researcher can be easily disproven using Shannon's sampling theorem.
Also, for his FAQ example of two peaks separated by 5 microseconds, my statement still holds true. The two peaks will "ring" the anti-alias filter in such a way that those two peaks still have an impact into the digital sampling (and on the eventual post-reconstruction-filter output). The waveform will look different and the high frequency spectra will be lost but the "information" of the two separate peaks will remain in the final analog output.
To be accurate and clear, the bit you can hear remains - which is all anyone wanting to listen to the audio signal should care about.
Cheers,
David.
They are supposed to repesentation two sounds occuring nearly, but not quite, simultaneously, for example two drums where the mallets hit 5 microseconds apart.
The time signals of two drums is not two impulses 5 microseconds apart. An impulse would normally represent something physical and have a known "height" and "width" to use Dr. K.s terminology (or was it vertical and horizontal?). What are the "heights" and "widths" in this case? If they are not known then what can you say about what should be going on in frequency space? Is "energy" conserved which is normally the first check performed when mapping between spaces?
Or are we talking about 2 periodic signals 5 microseconds apart? If so, what are they...
Unless there is a comprehensible statement of a problem one cannot say if it is right or wrong. All one can say is that it is incomprehensible.
The proper way to test would be [...]
Perhaps.
I just had a random thought about this and I wanted to run it by here:
The experimental procedure calls for a low-pass filtering of the signal. Does low-pass filtering significantly affect the perceived volume? Could the low-pass filter's effect on the perceived loudness be the perceived trait here? It is well-documented that decrease in loudness is perceived as decrease in quality. By slightly amplifying the filtered signal, could the perceived difference disappear?
So then, could the simple existence of the higher-order harmonics affect the overall power-spectrum (I'm aware that it will obviously be different in the frequency domain) significantly enough to be perceptible?
Well, the 4.7us time constant does decrease the level of the 7 kHz fundamental by 0.18 dB, which he mentions in his paper. According to a description in his paper of another experiment (done by a different author), the just noticeable difference is 0.7 dB at that frequency.
Please re-read the part I just bolded in the "quote" block above.
Agree. Sloppy reading on my part.
Then again, perhaps you should be a bit more explicit for those not versed in the art?
Here's another thing I noticed in his test apparatus. He describes the switching of the filter time constants and how it was optimized to minimize switching transients. He states that "...relays (which can induce a disturbance through magnetic flux change) were avoided...", although he does not say whether he tried any, and if so, whether this alleged disturbance was a problem in practice. The advantage of a relay would be to allow computer control of the switching, which when performed with the appropriate software, would allow for double-blind testing. IOW, it would allow for a test in which only the software, and not the test administrator, "knows" the state of the switch at any given time. He further states "...switch-bounce distortion was avoided by minimizing contact areas...", so this implies that a mechanical switch was used. Since that switch isn't a relay, it's just going to be a purely manual switching arrangement done by the test administrator physically flipping the switch. This in turn says that the test administrator is aware of the state of the switch at all times. It is therefore not a double-blind test. He does take care to call his tests "blind tests" rather than double-blind, so that is good. He uses the phrase "very effective blind listening tests" at one point in his article in fact. But he doesn't spell out what precautions, if any, were used to prevent the test subjects from observing the test administrator, which has the potential to influence the test results (AKA the Clever Hans Effect (http://en.wikipedia.org/wiki/Clever_Hans)).
Please re-read the part I just bolded in the "quote" block above.
Agree. Sloppy reading on my part.
Then again, perhaps you should be a bit more explicit for those not versed in the art?
When I get some spare time this week, I'll put together a good example including waveform and spectral plots. The basic idea is that the the magnitude and (more importantly) phase response of the digital signal will be the same below the nyquist frequency. The digital waveform might not look the same, but the information that your ears use to detect that 5 microsecond timing difference is still there.
we don't have to re-read all all of the time?
Further, there already is something interesting in an earlier
version of the FAQ document that is missing in the present one, if I remember
correctly.
It had to do with the time accuracy of 44.1kHz sampled systems ;-)
I wouldn't speculate. Don't speak such things unless you know them to be factual..that's how rumors get started, you know.
and no, he isn't updating his FAQ ie changing his answers... he is updating IE ACCOMMODATING NEW QUESTIONS THAT HE RECEIVES AFTER THE OLD FAQ WAS PUBLISHED>...
Here's another thing I noticed in his test apparatus. He describes the switching of the filter time constants and how it was optimized to minimize switching transients. He states that "...relays (which can induce a disturbance through magnetic flux change) were avoided...", although he does not say whether he tried any, and if so, whether this alleged disturbance was a problem in practice. The advantage of a relay would be to allow computer control of the switching, which when performed with the appropriate software, would allow for double-blind testing.
As a rule, relays don't induce significant disturbances in the signals they control when they change state. For example, one of the benchmarks for the ABX RM-2 relay module was an audibly transient-free switchover between two integrated amplifiers including phono or microphone input stages.
http://www.provide.net/~djcarlst/abx_rm2.htm (http://www.provide.net/~djcarlst/abx_rm2.htm)
The production RM2 was based on reed relays with 5 volt coils, but its first adequate prototype (equally effective) was based on traditional open-frame, open-contact relays with 28 volt coils. Damping networks were placed across relay coils to prevent arcing and damage to the control circuits.
Transient free switching of audio gear is not always trivial, but the important issues are related to the actual oder and kind (NC or NO) of contact openings. The original Clark JAES article tells all.
Paranoia about relay contacts is just another high end audiophile straw man. Professional audio gear and communications equipment has been reliably switching audio signals with relays for over a century.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
Perhaps you might detail the testing setup and apparatus, and the length of the test for us?
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
Perhaps you might detail the testing setup and apparatus, and the length of the test for us?
The links to the study are on that page, but here they are directly:
http://www.ocf.berkeley.edu/~ashon/a...ltrasonics.htm (http://www.ocf.berkeley.edu/~ashon/a...ltrasonics.htm)
http://jn.physiology.org/cgi/reprint/83/6/3548 (http://jn.physiology.org/cgi/reprint/83/6/3548)
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
In the opinon of many, no.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
Perhaps you might detail the testing setup and apparatus, and the length of the test for us?
The links to the study are on that page, but here they are directly:
http://www.ocf.berkeley.edu/~ashon/a...ltrasonics.htm (http://www.ocf.berkeley.edu/~ashon/a...ltrasonics.htm)
This link appears to be broken.
The correct link appears to be:
http://www.ocf.berkeley.edu/~ashon/audio/Ultrasonics.htm (http://www.ocf.berkeley.edu/~ashon/audio/Ultrasonics.htm)
This appears to be a year 2000 summary of the well-known Boyk paper that only documents the allready well-known existence of ultrasonic harmonics from musical instruments.
"This completely different paradigm for digital audio recording (SACD) does away with the anti-alias filters needed for PCM (CD and DVD-A) analogue waveform reconstruction. Preliminary reports in the audiophile community indicate that SACD has a natural quality of sound that DVD-A has yet to demonstrate. Ironically, one of the descriptions of SACD, this radically new digital format, is that it sounds "like analogue", meaning like LPs, ancient technology. (LPs i.e. vinyl records are associated with having smooth, relaxing presentations, that while sometimes not as impressive per audiophile standards, nevertheless can offer perfectly enjoyable music. The same is not always true, and in fact is seldom true, for CDs.)"
This is not evidence, it is supposition. Furthermore, the article contains errors of fact, some that were known to be errors when it was written, others that have come to light since it was written.
http://jn.physiology.org/cgi/reprint/83/6/3548 (http://jn.physiology.org/cgi/reprint/83/6/3548)
This is the oft-debunked Oohashi paper.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
In the opinon of many, no.
Opinion isn't good enough. The paper is scientific and peer-reviewed. It may be flawed, but if so it needs to be debunked scientifically. Can you provide links to similarly peer-reviewed refutations of the paper?
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Does this count?:
http://www.avsforum.com/avs-vb/showpost.ph...mp;postcount=28 (http://www.avsforum.com/avs-vb/showpost.php?p=17830969&postcount=28)
In the opinon of many, no.
Opinion isn't good enough.
Then your opinion is also not good enough, and that should be the end of your contributions to this thread. Do you want to go there? ;-)
The paper is scientific and peer-reviewed.
That does not make it infallible.
It may be flawed, but if so it needs to be debunked scientifically.
That has been done. The paper has always been weak. It was shopped around as a conference paper in several places before a journal that was weak enough to publish it was found.
Can you provide links to similarly peer-reviewed refutations of the paper?
That's unecessary. The universe is full of of peer-reviewed papers that became false when other relevant information became known. That is the nature of science.
Opinion isn't good enough. The paper is scientific and peer-reviewed.
So, if a paper comes out that claims, oh, say, that ESP works, and is published in an ESP journal, "opinion is not enough" even though there has never, ever, to date, been an unbiased test of ESP that shows it works. (Hint: Such papers exist.)
So, if a paper comes out in the journal of the ICR, showing that we were all created 6000 years ago, and the Grand Canyon was made during the "great flood", "opinion is not enough" even though in reputable journals there are thousands of pages on the erosion history of the Grand Canyon during the rise of the Colorado Plateau. (Hint: Such papers exist.)
So, if a paper comes out and gets published in a very odd journal that says "rocks fall up", "opinion is not enough" even if the rock always falls on your toe? (I'm not aware of any such papers, thankfully, although some TM practitioners claim you can levitate by meditation, I don't think even TM journals have published it as a "paper".)
It may be flawed, but if so it needs to be debunked scientifically. Can you provide links to similarly peer-reviewed refutations of the paper?
I think it's time for a reality check. "peer-reviewed refutations" is not what you will see in a journal. You're asking for something that is never going to exist, it's not how the system works.
What you will see is work showing something that reads on the subject, positive, negative, whatever.
And there is an entire forest of paper on that, dating from the 1920's to present.
Before any weight is put on the paper, it has to be REPEATED AND CONFIRMED. That is the scientific process. Going on about "peer reviewed" misses the mark in so many ways I hardly know where to start. The paper provides a presumed means to allow confirmation or perhaps not confirmation. As such, even if the reviewer has doubts, it may be published simply to start a dialog. Controversial papers are published, I ought to know, I published a couple of early papers on audio coding in light of reviewers who basically accused me of lying through my teeth. Yet, the papers were published, others also published, and now most everyone on the planet uses some form of the technology.
If you want to support this premise, repeat the experiment and see if you can confirm the results. You might even try to improve the experimental process, and try several different kinds of ultrasonic stimulii to see what's going on.
Finally, as a comment on ultrasonic stimulii, please check http://www.csgnetwork.com/atmossndabsorbcalc.html (http://www.csgnetwork.com/atmossndabsorbcalc.html) for the performance of the atmosphere under conditions of no wind, static atmosphere, etc.
Then consider the effects of wind and turbulence.
Before any weight is put on the paper, it has to be REPEATED AND CONFIRMED. That is the scientific process. Going on about "peer reviewed" misses the mark in so many ways I hardly know where to start. The paper provides a presumed means to allow confirmation or perhaps not confirmation. As such, even if the reviewer has doubts, it may be published simply to start a dialog. Controversial papers are published, I ought to know, I published a couple of early papers on audio coding in light of reviewers who basically accused me of lying through my teeth. Yet, the papers were published, others also published, and now most everyone on the planet uses some form of the technology.
The Oohashi paper is now about 10 years old. No confirmation, only failed attempts to confirm it.
Confirming it 10, 8 years ago (been there done that) was more difficult. Confirming it today is very easy. Interesting that so many of its advocates have not reported any of their own personal attempts to confirm it.
If you want to support this premise, repeat the experiment and see if you can confirm the results. You might even try to improve the experimental process, and try several different kinds of ultrasonic stimulii to see what's going on.
Been there, done that and I also failed to confirm it. Multiple test systems, multiple test music sources, multiple listeners.
Opinion isn't good enough. The paper is scientific and peer-reviewed. It may be flawed, but if so it needs to be debunked scientifically. Can you provide links to similarly peer-reviewed refutations of the paper?
You could have searched HA for previous discussion of Oohashi et al., you know. It's not like it hasn't been offered as evidence a half dozen times already.
http://www.hydrogenaudio.org/forums/index....st&p=606769 (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=68196&view=findpost&p=606769)
http://www.hydrogenaudio.org/forums/index....st&p=607276 (http://www.hydrogenaudio.org/forums/index.php?s=&showtopic=68348&view=findpost&p=607276)
Hello, I've just read the first and third articles of Milind Kunchur. I think that I found a weak point in his work. But first, a few comments on the discussion
Since changes in this second harmonic value could explain the experimental results without needing some hypothesis regarding the alleged ability of the ear to detect signals above the frequency limit of human hearing, it's essential to include these data. Yet he fails to do so.
Kunchur speaks about second harmonics created in the middle ear, after the eardrum, but before reaching the hair cells, where ultrasonic frequencies are dead stopped. Therefore the hypothesis remains correct even of no second harmonic at all is outputted from the speaker.
However, he doesn't speak about the amplitude of this harmonic, probably unknown, and let an arbitrary factor b in front of the calculus.
In practice, this intermodulation seem to be extremely low, if it ever exists, as 2bdecided recalls :
all previous proper experiments show that, apart from via bone conduction, ultrasonics don't distort to create audible frequencies in the ear.
In equipment, yes, but not in the ear.
If someone proves otherwise (quite possible), it'll be interesting.
IIRC there was someone here who did (playing ultrasonic from a separate audio system(!) and ABXing presence / absence in the presence of an audible sound), but I didn't see how the thread ended.
I remember that Nika Aldrich and I, in an old discussion in George Massenbug forum, once ran this experiment : play a set of two frequencies in a speaker, one audible, one inaudible. I chose 12 kHz and 18 kHz. The 4 kHz intermodulation was clearly audible. Then, play one of them in the left speaker of your stereo, and the other in the right speaker. No intermodulation was audible anymore.
From the first paragraph at: www.physics.sc.edu/kunchur/Acoustics-papers.htm (http://www.physics.sc.edu/kunchur/Acoustics-papers.htm)
Our recent behavioral studies on human subjects proved that humans can discern timing alterations on a 5 microsecond time scale, indicating that that digital sampling rates used in common consumer audio (such as CD) are insufficient for fully preserving transparency.
This statement contains a glaring error:
44.1 kHz sampling PCM systems are perfectly capable of reproducing the phase of audible frequencies to picosecond accuracy. The need for 5-microsecond temporal accuracy does NOT indicate the need for a higher sample rate. It simply indicates that jitter must be less than 5-microseconds.
True, but Kunchur's first paper is actually not about temporal resolution. The blind test was comparing a full-band signal versus a lowpassed version of the same signal. Thus the conclusion about sample rates is relevant.
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Kunchur's results
are positive !
We can question them, critic the methodology, find flaws and discuss them, so the conclusion is not clear yet, but they are positive.
Well, the 4.7us time constant does decrease the level of the 7 kHz fundamental by 0.18 dB, which he mentions in his paper. According to a description in his paper of another experiment (done by a different author), the just noticeable difference is 0.7 dB at that frequency.
...and that's the weak point that I wanted to talk about. The 0.7 dB threshold is assumed without checking. Kunchur says in the FAQ that he can't redo everything and that evaluating again the Just Noticeable Difference would take about two years.
However, since this point annoyed me, I ran an ABX test between two 7 kHz sines. I chose the level difference that Kunchur got for the 5.6 µs lowpass experiment : 0.25 dB.
I generated two 7 kHz sines using Soundforge. Since the software is all 16 bits, I decreased the volume of the first one by 1 dB, and of the second one by 1.25 dB, in order to get the same quantization noise on both.
I could ABX them with a score of 8/8 with Foobar2000 ABX module, no DSP, no replaygain. I was wearing headphones and the playback volume was moderate. Inferior to 80 dB, but I couldn't say how much.
I had to perform instant switches back and forth in order to hear the difference. Foobar2000 cuts the signal for a fraction of second each time I press a button. Like the listeners in Kunchur's experiments, I found the louder sine "brighter".
Therefore the assumption that the level Just Noticeable Difference for a 7 kHz sine at 69 dB is 0.7 dB seems seriously flawed, and that questions all Kunchur's results.
Therefore the assumption that the level Just Noticeable Difference for a 7 kHz sine at 69 dB is 0.7 dB seems seriously flawed, and that questions all Kunchur's results.
Pages 180 and 181 of "Psychoacoustics: Facts and Models" by Fastl/Zwicker say it all:
http://books.google.com/books?id=eGcfn9ddR...1&lpg=PA181 (http://books.google.com/books?id=eGcfn9ddRhcC&pg=PA181&lpg=PA181)
[
Edit: http://www.google.com/search?q=just-notice...evel+difference (http://www.google.com/search?q=just-noticeable+level+difference), in case the above link doesn't work]
A typical characteristic of just-noticeable amplitude modulation of sinusoidal tones is the level dependence. Almost the same level dependence is measured for just-noticeable differences in level [JNDL], as indicated in Fig. 7.5. For low sound pressure levels below 20 dB, JNDL increases greatly towards threshold, but decreases from about 0.4 dB at 40 dB to about 0.2 dB at 100 dB sound pressure level. The decrement seems to be not quite as strong as for amplitude modulation, but the characteristic is similar to the data shown by the solid line in Fig. 7.1. This characteristic is almost independent of frequency if, instead of the sound pressure level, the level above threshold (or, even better, the loudness level) is used as the abscissa.
A 7-kHz tone is only a few dB quieter than a 1-kHz tone: http://en.wikipedia.org/wiki/Equal-loudness_contour (http://en.wikipedia.org/wiki/Equal-loudness_contour). So with your 0.25 dB, you're well in line with Zwicker's findings.
Chris
Nobody seems to be able to do it. Instead, all we get are these abstract tests like Kunchur's, involving some most definately non-musical waveforms with questionable relevance and inconclusive results.
I agree. Not a single positive result in a decade is simple but mortgageable evidence.
Kunchur's results are positive !
We can question them, critic the methodology, find flaws and discuss them, so the conclusion is not clear yet, but they are positive.
Well, the 4.7us time constant does decrease the level of the 7 kHz fundamental by 0.18 dB, which he mentions in his paper. According to a description in his paper of another experiment (done by a different author), the just noticeable difference is 0.7 dB at that frequency.
...and that's the weak point that I wanted to talk about. The 0.7 dB threshold is assumed without checking. Kunchur says in the FAQ that he can't redo everything and that evaluating again the Just Noticeable Difference would take about two years.
However, since this point annoyed me, I ran an ABX test between two 7 kHz sines. I chose the level difference that Kunchur got for the 5.6 µs lowpass experiment : 0.25 dB.
I generated two 7 kHz sines using Soundforge. Since the software is all 16 bits, I decreased the volume of the first one by 1 dB, and of the second one by 1.25 dB, in order to get the same quantization noise on both.
I could ABX them with a score of 8/8 with Foobar2000 ABX module, no DSP, no replaygain. I was wearing headphones and the playback volume was moderate. Inferior to 80 dB, but I couldn't say how much.
I had to perform instant switches back and forth in order to hear the difference. Foobar2000 cuts the signal for a fraction of second each time I press a button. Like the listeners in Kunchur's experiments, I found the louder sine "brighter".
Therefore the assumption that the level Just Noticeable Difference for a 7 kHz sine at 69 dB is 0.7 dB seems seriously flawed, and that questions all Kunchur's results.
Disclaimer: My comments were about musical sounds, not lengthy pure sine waves.
It is interesting to note that now the same people at stereophile are ragging on Sean Olive. First a guy claimed that Sean's blog was sending out viruses, and that got squashed, now he's claiming that Sean doesn't know how to run a listening test.
This is the same person who alleges I don't know what an impulse response is, doesn't know the difference between an impulse response and signal detection, and who regards my speaking on a panel with Ethan Winer as "travelling 6000 miles to aid a competitor".
It is interesting to note that now the same people at stereophile are ragging on Sean Olive. First a guy claimed that Sean's blog was sending out viruses, and that got squashed, now he's claiming that Sean doesn't know how to run a listening test.
This is the same person who alleges I don't know what an impulse response is, doesn't know the difference between an impulse response and signal detection, and who regards my speaking on a panel with Ethan Winer as "travelling 6000 miles to aid a competitor".
I presume this is the Stereophile forum, not the ragazine. I've stopped paying much attention to what happens there, because it really doesn't seem to matter a lot in the real world.
It is interesting to note that now the same people at stereophile are ragging on Sean Olive. First a guy claimed that Sean's blog was sending out viruses, and that got squashed, now he's claiming that Sean doesn't know how to run a listening test.
This is the same person who alleges I don't know what an impulse response is, doesn't know the difference between an impulse response and signal detection, and who regards my speaking on a panel with Ethan Winer as "travelling 6000 miles to aid a competitor".
You're not doing it right JJ, Stereophile will not even let me join the forum. It is just great, I can't go crazy trying to straighten those BOZO posters out on that forum.
Paul
It is interesting to note that now the same people at stereophile are ragging on Sean Olive. First a guy claimed that Sean's blog was sending out viruses, and that got squashed, now he's claiming that Sean doesn't know how to run a listening test.
This is the same person who alleges I don't know what an impulse response is, doesn't know the difference between an impulse response and signal detection, and who regards my speaking on a panel with Ethan Winer as "travelling 6000 miles to aid a competitor".
I presume this is the Stereophile forum, not the ragazine. I've stopped paying much attention to what happens there, because it really doesn't seem to matter a lot in the real world.
This is true, they've basically allowed the absolutely crazy and/or willfully dissembling types to drive everyone with much sense right off the board. The only one I know they banned was Ethan, and they banned one of the worse crazies along with him, but somebody who writes in the style of the now-missing Alan Derridia is not much loss in my book.
It is interesting to note that now the same people at stereophile are ragging on Sean Olive. First a guy claimed that Sean's blog was sending out viruses, and that got squashed, now he's claiming that Sean doesn't know how to run a listening test.
This is the same person who alleges I don't know what an impulse response is, doesn't know the difference between an impulse response and signal detection, and who regards my speaking on a panel with Ethan Winer as "travelling 6000 miles to aid a competitor".
I presume this is the Stereophile forum, not the ragazine. I've stopped paying much attention to what happens there, because it really doesn't seem to matter a lot in the real world.
This is true, they've basically allowed the absolutely crazy and/or willfully dissembling types to drive everyone with much sense right off the board. The only one I know they banned was Ethan, and they banned one of the worse crazies along with him, but somebody who writes in the style of the now-missing Alan Derridia is not much loss in my book.
Tit-for-tat paired banning seems be the rage. I think that happened to me at the Womb. I made only 2 fairly innocent posts over and was banned for life with no explanation about the same time one of theirs was being moderated over here.
Of course, these conferences are usually bannning someone who posts under an alias, along with someone like Ethan who posts only under his legal name. The anonymous poster is back the next day under a new alias and that is that.
I don't think that either forum is about any basic search for truth.
I've got more important projects languishing for the time to complete them...
I've had the good fortune (or timid personality depending on how you look at things) to avoid getting banned anywhere.
As long as only audiophiles give a hoot about the Stereophile forums, I'm not sure I can really care about the constant skullduggery I hear emanating from there. I mean, if it had as much cachet as Head-Fi does for headphones - and I have repeatedly referred people to Head-Fi to research headphone purchases - or the less crazy vinyl sites, I'd care a lot more. But that forum just doesn't have any rep whatsoever AFAIK. There isn't a whole lot of difference between "not criticized by anybody" and "criticized but only by crazies".
Now, that is in stark contrast to the original topic (the Kunchur papers) and those are, for the same reasons, a much bigger deal.
these conferences are usually banning someone who posts under an alias, along with someone like Ethan who posts only under his legal name. The anonymous poster is back the next day under a new alias and that is that.
Indeed. The Frog is back under a new name. Advising the authorities had no effect, because apparently it's only me they needed to get out of their forum.
--Ethan
these conferences are usually banning someone who posts under an alias, along with someone like Ethan who posts only under his legal name. The anonymous poster is back the next day under a new alias and that is that.
Indeed. The Frog is back under a new name. Advising the authorities had no effect, because apparently it's only me they needed to get out of their forum.
--Ethan
I thought so too, but seems more prone to arguing woowoo science and less prone to ad-homs... A bit, at least.
But, well, nothing I can say would be anything but disparaging to the new arrival, so say nothing substantitive I shall.
we don't have to re-read all all of the time?
Further, there already is something interesting in an earlier
version of the FAQ document that is missing in the present one, if I remember
correctly.
It had to do with the time accuracy of 44.1kHz sampled systems ;-)
I wouldn't speculate. Don't speak such things unless you know them to be factual..that's how rumors get started, you know.
and no, he isn't updating his FAQ ie changing his answers... he is updating IE ACCOMMODATING NEW QUESTIONS THAT HE RECEIVES AFTER THE OLD FAQ WAS PUBLISHED>...
Apparently he's not doing either.
http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm (http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm)
FAQs (this document answers FAQs, and provides some tutorial information for the above papers), M. N. Kunchur, web article (last updated July 15, 2009).
Cursory glance around the Net suggests that 'hi rez' proponents already consider his work as 'proof' that Redbook rate sampling is both intrinsically audible and degradative to sound.
His webpage is rather amusing. As one of his 'publications' he links to a .pdf (http://www.physics.sc.edu/~kunchur/papers/Music-Psychology-book-pages.pdf) of a photo of a page from a textbook that cites (uncritically, and as it happens incorrectly) his 2008 paper's claim about temporal resolution of human hearing.
Now that's an ancient thread bump I can approve of, krabapple!
I recently revisited Monty's excellent recent video, and noticed that he debunks Kunchur soundly using a very clear example with no math:
Monty Montgomery explains digital audio (http://xiph.org/video/vid2.shtml)
That part starts near the end at around 21:55.
--Ethan
That video could clear up so many misconceptions newbies have.
It's really very simple to disprove such claims. All one has to do is sample a 20kHz sine wave (with no dithering to make things clearer to the uninformed with two phases 5 degrees apart.
Measure the difference in the samples. Oops, there is a sine wave encoded in that difference, just like you'd expect.
That is absolute, concrete evidence that Kuncher, Fraboni, and a bunch of other people talking about the "dirty little secret of digital audio" are utterly full of cow patties.
Of course, even the IEEE published an article of Fraboni's in which he made that claim about the "dirty little secret". His claim, of course, is utterly false, and the Spectrum publishing that bit of quackery is just an astonishing violation of the IEEE's own ethical code.
They were pressured by a variety of people to retract that, and blew it off completely, claiming it was a "controversy".
Apparently he's not doing either.
http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm (http://www.physics.sc.edu/~kunchur/Acoustics-papers.htm)
I was interested to see how he would handled the position he had put himself in. I cannot believe he did not look up the basics of sampling after the criticism he received which makes the 2010 AES presentation interesting. He hasn't kept his head down completely. Has anyone looked closely enough at the slides to determine how much of the original nonsense is retained?
are utterly full of cow patties.
Oh JJ, you're too kind.
are utterly full of cow patties.
Oh JJ, you're too kind.
Should I have said "udderly full of cow patties"?