MQA creator Bob Stuart answers questions.

[nscrivener]

The calls to submit himself to the blind tests to prove his claim conveniently coincide with this sudden reality-barrier to have anymore time to further continue discourse.

Curious at the least, but raises eyebrows.

I sometimes feel like folks here focus at times, on the technical aspects of claims, that they aren't seeing that the whole ordeal can be finished quite quickly by proposing the claimant demonstrate extraordinary claims. Or instances where the claims are tested using sound reasoning and basic scientific methods devoid of eyebrow raising caveats with studies that are plagued with issues, and/or conflicts of interest.

The guy himself knows more than I do even on the topics of electronics without a doubt.

Where he fails, is in his basic reasoning and behavior.. (as you can see between the limited exchange we've had).

For me, it became clear that he had an agenda when he listed his nonsensical list of 'known phenomena' that 'don't fit the existing model', which anyone with a background in experimental physics could easily see have many possible explanations and therefore don't in any way prove the inadequacy of the model. Then he went on to claim 'insight' that he chose not to share because it might reveal something that the MQA people have diligently tried to conceal!

I don't think that it's plausible that somebody with his background and intellect would behave in that manner unless there was an agenda. He was here to sow doubt amongst the members here. That's the only reasonable conclusion.

 

[SIY]

I don't think that it's plausible that somebody with his background and intellect would behave in that manner unless there was an agenda. He was here to sow doubt amongst the members here. That's the only reasonable conclusion.

A more elegant and polite way to say the same thing I did.

 

[LTig]

You are describing the process of conversion of a bandlimited signal into digital values.

Bob Stuart and me were referring to the process of conversion of original signal into a bandlimited one. How is that accomplished? Classically, by averaging of some sort: for instance employing a capacitor in a classic low pass filter.

If you look at the digital low pass filters, they employ averaging as well. Well, in the digital domain you can do decimation instead of averaging, but then you are throwing away potentially useful information, and are less flexible with the values of cutoff frequency that you can use.

Bandlimiting before sampling must be done analog, of course. Bandlimiting in the digital domain is only required for decimation/downsampling, the digitized signal must already be bandlimited to less than half of its current samplerate.

Figure 11 shows an analog signal and below a few samples. As one sees clearly the impulse is much shorter then one sample and therefor its frequency range is much higher than the samplerate. Hence after proper analog bandlimiting its shape must be so broad that it covers more than 2 samples. I would expect that it is just zero, and then its position in time is meaningless.

BTW analog filtering is done with more sophisticated filters than just a simple cap. Simple averaging is quite a bad filter. This is especially true for digital filters. They are usually implemented as IIR or FIR filter.

ADCs for audio usually work internally with much higher samplerates to ease the design of the antialiasing filter. After sampling a digital filter is used to bandlimit the captured samples to the desired output sample rate, and the resulting sample stream is then decimated and clocked out.

 

[LTig]

If the signal coming from the microphone is naturally bandlimited so that the faithful digitization at a given sampling rate is possible, we have nothing to argue about. The classic theory prescribes what needs to be done perfectly well. In that case I totally agree with you.

The more interesting discussion is about the situation when this is not the case, and for whatever reason we need to bandlimit the analog signal before digitizing at a target sample rate. The bandlimiting may be done in analog and/or digital domain.

Bandlimiting must always be done in the analog domain if the analog signal contains frequencies higher than half the sample rate of the ADC. Only if the ADC uses a very high initial samplerate one could omit the analog antialiasing filter and perform bandlimiting followed by decimation in the digital domain (this usually is the case, but I would still add an analog antialisaing filter to make sure that no aliasing occurs).

Whatever bandlimiting filter you decide to use, you will introduce artifacts. For instance, with FIRs you'll get pre- and pos-ringing; with IIR, you'll get potentially non-benign phase irregularities; and you'll get a bit of both, in smaller quantities, with hybrid filters. Analog circuits can be approximated by digital filters, using infinitesimally small timing steps, and exhibit presence of the same artifacts.

All test signals used to show the ringing of filters are artificially calculated and are not bandlimited, so the ringing can show its ugly head. If the signal is bandlimited (as in sampled audio) ringing does not occur. It's just that a bandlimited signal with artificially sharp steps looks like this after the higher frequencies have been removed (Gibbs phenomenon).

 

[restorer-john]

So we are resuming normal programming now that Sergie Bob has left the room?

 

[Blumlein 88]

I had a short stint in a gamelan ensemble when I was in college. Interesting sound, but I got bored quickly and went back to playing American blues.

Surely somewhere someone is doing gamelan blues???

 

[SIY]

Surely somewhere someone is doing gamelan blues???

Sorta.

 

[Sal1950]

Blues, is that what they call that? Ugg

Let me cue up some Bobby (Blue ) Bland

 

[SIY]

Not my cuppa tea either- we have tickets to see Popa Chubby in September, and that's more my speed.

 

[Sergei]

Unfortunately, no free lunch for the mostly older farts we are...

View attachment 27246

View attachment 27247

Yes, not pretty. But this is even less so (from http://www.cochlea.eu/en/pathology/presbycusis):

Apologies for not digging up source articles: busy time. From the top of my head, a healthy elderly man by 70 y.o. shall expect to have a loss of about -70dB at 7 KHz, loss of about 70% of frequency selectivity, yet keep over 70% of ITD precision.

I wish we could live our lives on a quiet island. It helps (from http://www.psych.usyd.edu.au/staff/alexh/teaching/auditoryTute_2016/):

 

[Sergei]

(A) - Higher sample and bit rates are required in recording due to the extensive processing and mixing that occurs. It is a form of headroom. It's unclear why studios NOT deeming it necessary to use even higher sample/bit rates than 192/24 favours your argument?.

I believe it does. Why 192? What's so special about this number? Do you have other theory explaining this?

(B) - Expectation bias? The recording/mastering is not as good as you thought? Your CD playback mechanism is somehow audibly inferior to your SACD playback mechanism?

(C) - Expectation bias? Personal taste?

I wasn't satisfied with these kinds of partial explanation. Wanted to find something that explains the maximum number of strange and controversial phenomena of the audiophile world.

The trouble with your proposed 'new paradigm' is that it doesn't make sense. You've already been provided information showing why temporal resolution is not related to sample rate.

I'm thinking about how to demonstrate to the members of this forum, in a a simple and reproducible way, that it is in fact is related - in the hearing system that is. I agree that in the idealized DSP domain it isn't related much.

 

[Sergei]

A sample rate of 192 kHz is thus ample for capturing the signal perfectly.

Did I ever argue otherwise?

Are you under NDA with MQA?

No. Neither with Meridian or Algol. If I were, why would I post here?

 

[Krunok]

So just to have a safety factor it seems to me if you can deliver the signal to the speaker or headphone with no more than .01% distortion and with flat frequency response you are golden.

Sounds reasonable. Although, to be honest, even when fast switching my tube amp with 0.2% THD and Rotel amp with 0.004% THD I can only tell them apart because of tube amp HF roll off but not because of difference in distortion.

 

[Sergei]

You went around in a circle though. You posted a study he essentially dismantled three years ago..

Did he? Could you post a link please? I've only seen what Amir wrote in
https://www.audiosciencereview.com/forum/index.php?threads/statistics-of-abx-testing.170/
and comments on the 2014 study
https://secure.aes.org/forum/pubs/conventions/?ID=416

Did I read those incorrectly? Amir appeared to be supportive of the study.

Also, it's not pointless to fulfill the request if you honestly undertake it with the basic parameters. Also why do you care if it is dismissed by people as long as you fulfilled their request. You can do yourself a favor and at least dismiss all criticisms that can be leveled against you like:

"So you're just going to talk about this, and not prove it yourself?"

This way, no third-party outsiders can accuse you of laziness, or fearing you were hiding behind claims you couldn't back up or didn't want to subject yourself to.

So no, it is not pointless, and will at the very least demonstrate an air of good faith if nothing else at the least. If that is pointless to you, then we are eons apart with respect to our value systems (though I already suspect this seeing how you're willing to link a study that has glaring flaws), especially considering you're willing to stand behind such studies yourself.

I accept the "good faith" argument. Didn't realize it was some kind of "rite of passage" on this forum. Thanks for explaining.

Let me explain in turn why I don't believe it would be a particularly illuminating exercise.

15-10 years ago I was big time into something similar as an active audiophile: comparing CD and SACD sound quality, telling others that I can hear the difference, hearing back that it must be confirmation bias, or flaws in equipment, or different mixes for CD (bad) and SACD (good) etc. - pretty much the same arguments I'm getting on this forum from some of the members.

I was buying ever-better equipment, more CDs, more SACDs, was visiting other audiophiles houses with my CDs and SACD to test on their equipment, and so on. So I'm kind of chuckle hearing that I may be perceived lazy in pursuing the "truth" in this subject area, especially remembering the month during which I tested several high-end A-class amplifiers in the row: hauling and installing those iron behemoths was quite an endeavor.

After a while, I realized that I could not indeed hear the difference if the music played wasn't sophisticated enough, especially if it didn't have many transients. As I already mentioned on this forum, four musicians or less playing traditional instruments, or singing, would make the CD and SACD versions sound equivalent to me.

I consider the "Joseph Haydn: String Quartet In D, Op. 76, No. 5 - Finale - Presto" to be a simple music piece, and I don't expect to hear any difference between the versions with different sampling rates available from 2L. When I have time, I'm going to find out whether they have something more sophisticated.

I also realized back then that some other people genuinely can't hear the difference, even on sophisticated pieces. And some yet other people claimed, but I didn't get to meet them in person unfortunately, that they can hear the difference irrespective of the type of music, as long as it was a commercial record rather than speech or test tones.

Indeed, nowadays it is easier to do the testing, with hi-res samples available for download. And I did this too, about six years ago if memory serves me. The result was that for half a dozen moderately sophisticated music pieces I could tell the difference between 44/16 and 192/24 better than by chance, but couldn't reliably distinguish either from 96/24.

Also the study has an air of source bias that cannot be overlooked by rational people (I didn't want to bring this us, but it had to be done by someone).

A more comprehensive meta-study (http://www.aes.org/e-lib/browse.cfm?elib=18296), combining 400 participants in over 12,500 trials reveals that there is indeed statistically significant difference, yet it is small, 5% at most.

With such extensive work already done by the researchers, I don't feel it makes much sense to me to participate in redoing studies of this type on a smaller scale.

For me, as of today, it is much more interesting to understand why the difference can be detected, and why the detection capability differs so much in different people.

Interestingly enough, studies implicating the role of timing identification precision differences appeared as early as mid-1990s, if not earlier. For instance: http://www.aes.org/tmpFiles/elib/20190606/7217.pdf.

But still, even today on this forum, people don't believe me when I discuss that. I guess the full generation cycle is not over yet. 1997 + 25 = 2022. Let's wait and see ...

In conclusion I also wanted to draw attention to you not addressing or conceding the point where I exclaim it would be ridiculous to provide you with the information you asked for prior. Like explanations of what degree disagreed upon methodologies have impacted the findings. I would appreciate it (also a form of good faith in discourse) if you can state your reply to this statement I make about the absurdity of someone going out and explaining to numerical degree what the "preposterousness level" of undertaking something of that nature is..

I don't enjoy putting people in awkward situations. I hoped that you'd re-read what I actually wrote and drop you inquiry. But if you insist, that's what I wrote:

"Instead of repeating the study with higher precision, and either confirming or refuting its results, they started attacking the study methods. Even if a method was flawed, it is still unclear to what degree this affected the result. So, in the minds of these attackers, they had their "proof" that the study is invalid, without doing the legwork."

The meta-study I referred to above, done in 2016, a year and a half after the 2014 study, is an example of such "legwork", done with higher precision. It didn't overturn the results of the 2014 study, yet could have. That's the right way to do it in science.

I hope we are square now. Peace?

 

[Sergei]

This I will never understand. All this talk about hearing supra 22Khz, yet everyone I have seen talk about mics used for recording music stop around that threshold of audibility.

Why are any of these talks AT ALL entertained considering these mics aren't recording any of those frequencies? Unless I am missing something here like the possibility that music recording studios have multiple mics?

Some recording as they do now, and others recording Ghz frequencies or something idiotic like that "just to be safe".. I'd bet this is purely as ridiculous as I think it to be.

So again I ask, why is all this "audibility of ultrasonics" EVER a discussion point if we're talking about recorded music?

While not actual measurements, the graphs below may be taken as illustrations of microphones reaction to a shockwave (from https://earthworksaudio.com/support/technology/impulse-response/). Notice straight, instead of sinusoid-like, fragments of the graph.

Shockwave is (from https://en.wikipedia.org/wiki/Shock_wave):

"a type of propagating disturbance that moves faster than the local speed of sound in the medium. Like an ordinary wave, a shock wave carries energy and can propagate through a medium but is characterized by an abrupt, nearly discontinuous, change in pressure, temperature, and density of the medium."

Basically, a bunch of air molecules gets thrown at the microphone diaphragm all at once. The diaphragm accelerates virtually instantaneously (notice the sharp upturn), then moves at a constant speed. If the microphone is of dynamic type, the nearly instantaneous transition from zero speed to a constant speed corresponds electronically to a good approximation of a step function.

When the diaphragm hits the limit of mechanical motion, it bounces back and keeps going at a constant, yet now slower speed. Electronically, this corresponds to a first half-cycle of a rectangular wave, which generates a nasty set of harmonics extending into ultrasound region, receding rather slowly: https://en.wikipedia.org/wiki/Square_wave.

The sources of shockwaves are more abundant in a rock band or an orchestra than one might think. For instance, inside a trombone: http://www.physics.mcgill.ca/~guymoore/ph225/shock.pdf. Over distance, a shockwave dissipates and transforms into regular harmonic waves: that's what the listeners are supposed to hear. Yet a microphone placed too close, or right on the trombone, may be affected by a shockwave.
A microphone placed very close to an electric guitar amplifier, playing a loud passage in overdrive, might also experience periodic shockwaves.

The ultrasonics can be stopped by an analog low pass filter of course. Sometimes microphone pre-amp has such a filter, sometimes not - in the interest of "sonic transparency" during operation in regular mode. The "transparency" could relate to lower noise and less significant frequency response and phase irregularities, inherent in filters.

Also, there is an area of shapes of air disturbances in between: a mix of shockwave front and arriving slightly later regular sinusoid, which different microphones may have very different and at times peculiar reactions to. Like: when the shockwave front and the sinusoid arrival happen to be spaced just so, the signal spike could be much higher.

Condenser microphones with well-designed electronics are usually more immune to shockwaves translating into weird outputs, yet their diaphragms are much lighter, and may thus move even quicker given the same mechanical momentum transferred by a shockwave. So, ultimately they may output a rounded shape which is neither square nor sinusoid: not as nasty harmonics-wise, yet not completely benign either.

I hope this answers your question about how a microphone, nominally rolling off at 20 KHz due to its mechanical design, can generate ultrasonics far up the spectrum in its electric output. Some recording engineers deal with this by using carefully designed, precisely made, and thus rather expensive microphone preamps. Others sample at very high frequencies and filter later in software - thus the seemingly unneeded "GHz" you mentioned.

 

[SIY]

Tough to sort through a Gish Gallop for picking apart misleading and irrelevant info, but here's a few bits:

The inside of a trombone is not terribly relevant. Outside the trombone, there's not much ultrasonic energy.

The Earthworks mikes with extended ultrasonic response are not very useful for recording because of noise. I have first hand experience with this. These are measurement mikes, not recording mikes.

Analog anti-aliasing filters are the norm, so that part is likewise irrelevant, as is the "might" "could" "possibly" scenarios pulled out of the air.

No mike I'm aware of is going to have the diaphragm slamming into the backplate unless the recording engineer is exceedingly stupid in his mike and positioning choice and doesn't bother noticing the gross overload during level setting. FWIW, 130-150 dB SPL is not atypical for mike overload point.

I could make this much longer, but as usual, a lot of handwaving nothing which really doesn't deserve response.

 

[nscrivener]

Did he? Could you post a link please? I've only seen what Amir wrote in
https://www.audiosciencereview.com/forum/index.php?threads/statistics-of-abx-testing.170/
and comments on the 2014 study
https://secure.aes.org/forum/pubs/conventions/?ID=416

Did I read those incorrectly? Amir appeared to be supportive of the study.



I accept the "good faith" argument. Didn't realize it was some kind of "rite of passage" on this forum. Thanks for explaining.

Let me explain in turn why I don't believe it would be a particularly illuminating exercise.

15-10 years ago I was big time into something similar as an active audiophile: comparing CD and SACD sound quality, telling others that I can hear the difference, hearing back that it must be confirmation bias, or flaws in equipment, or different mixes for CD (bad) and SACD (good) etc. - pretty much the same arguments I'm getting on this forum from some of the members.

I was buying ever-better equipment, more CDs, more SACDs, was visiting other audiophiles houses with my CDs and SACD to test on their equipment, and so on. So I'm kind of chuckle hearing that I may be perceived lazy in pursuing the "truth" in this subject area, especially remembering the month during which I tested several high-end A-class amplifiers in the row: hauling and installing those iron behemoths was quite an endeavor.

After a while, I realized that I could not indeed hear the difference if the music played wasn't sophisticated enough, especially if it didn't have many transients. As I already mentioned on this forum, four musicians or less playing traditional instruments, or singing, would make the CD and SACD versions sound equivalent to me.

I consider the "Joseph Haydn: String Quartet In D, Op. 76, No. 5 - Finale - Presto" to be a simple music piece, and I don't expect to hear any difference between the versions with different sampling rates available from 2L. When I have time, I'm going to find out whether they have something more sophisticated.

I also realized back then that some other people genuinely can't hear the difference, even on sophisticated pieces. And some yet other people claimed, but I didn't get to meet them in person unfortunately, that they can hear the difference irrespective of the type of music, as long as it was a commercial record rather than speech or test tones.

Indeed, nowadays it is easier to do the testing, with hi-res samples available for download. And I did this too, about six years ago if memory serves me. The result was that for half a dozen moderately sophisticated music pieces I could tell the difference between 44/16 and 192/24 better than by chance, but couldn't reliably distinguish either from 96/24.



A more comprehensive meta-study (http://www.aes.org/e-lib/browse.cfm?elib=18296), combining 400 participants in over 12,500 trials reveals that there is indeed statistically significant difference, yet it is small, 5% at most.

With such extensive work already done by the researchers, I don't feel it makes much sense to me to participate in redoing studies of this type on a smaller scale.

For me, as of today, it is much more interesting to understand why the difference can be detected, and why the detection capability differs so much in different people.

Interestingly enough, studies implicating the role of timing identification precision differences appeared as early as mid-1990s, if not earlier. For instance: http://www.aes.org/tmpFiles/elib/20190606/7217.pdf.

But still, even today on this forum, people don't believe me when I discuss that. I guess the full generation cycle is not over yet. 1997 + 25 = 2022. Let's wait and see ...



I don't enjoy putting people in awkward situations. I hoped that you'd re-read what I actually wrote and drop you inquiry. But if you insist, that's what I wrote:

"Instead of repeating the study with higher precision, and either confirming or refuting its results, they started attacking the study methods. Even if a method was flawed, it is still unclear to what degree this affected the result. So, in the minds of these attackers, they had their "proof" that the study is invalid, without doing the legwork."

The meta-study I referred to above, done in 2016, a year and a half after the 2014 study, is an example of such "legwork", done with higher precision. It didn't overturn the results of the 2014 study, yet could have. That's the right way to do it in science.

I hope we are square now. Peace?

It's fair to say that in carefully controlled conditions, some trained listeners can discriminate high resolution audio from standard resolution audio, on some material.

It's also fair to say that anti-aliasing filters can exhibit measurable characteristics of pre and post ringing when fed test signals that include very sharp transients (i.e. square waves).

Where I have a problem is the extent to which you draw conclusions from the above.

Even if I was to grant your premise, essentially the arguments that MQA have put forward regarding the perceptual effect of "smearing" in the time domain, which I don't, because I don't think that the reasons for the differences noted above have been established, at least not with any certainty, there is absolutely no doubt whatsoever that there are other factors influencing the quality of digital audio playback that are many orders of magnitude more important than the incremental improvements that might come from increased sample rates, bit rates, and different low-pass filtering schemes. Take a very well recorded and mastered track encoded in 16 bit 44khz PCM and compare it to an average track in any high resolution format. There is no doubt that the former will sound better. Possibly much better. I have many albums in my collection that demonstrate that. It's also true that factors like speaker selection and room acoustics have huge effects on playback quality, also orders of magnitude more significant. I'd almost go so far as to say that you could probably pick just about anything else in the playback chain to focus on and have a greater chance of getting audible improvements. (Ok maybe that's pushing the point a little too far.)

 

[SIY]

It's fair to say that in carefully controlled conditions, some trained listeners can discriminate high resolution audio from standard resolution audio, on some material.

It's also fair to say that anti-aliasing filters can exhibit measurable characteristics of pre and post ringing when fed test signals that include very sharp transients (i.e. square waves).

I'm not sure I'd go with either. Can you cite anything reliable on the first (assuming that by "standard resolution" you mean something like 16/44 or 16/48, and no diddling of volume knobs to bring up the noise floor during silent parts)?

Regarding the second, I'm at a loss as to how an analog filter can show preringing. Am I missing something?

 

[restorer-john]

Regarding the second, I'm at a loss as to how an analog filter can show preringing. Am I missing something?

Nope, just the pre-ringing.

Sony CDP-101, analogue LPF, single sample impulse, 0dBFS. Denon audio technical test disc.

 

[nscrivener]

I'm not sure I'd go with either. Can you cite anything reliable on the first (assuming that by "standard resolution" you mean something like 16/44 or 16/48, and no diddling of volume knobs to bring up the noise floor during silent parts)?

Regarding the second, I'm at a loss as to how an analog filter can show preringing. Am I missing something?

Regarding the first. Amir is of that opinion as contained in his following post: https://www.audiosciencereview.com/forum/index.php?threads/high-resolution-audio-does-it-matter.11/

Also, Sergei linked to the following metastudy:
http://www.aes.org/e-lib/browse.cfm?elib=18296

Does that not look reputable to you?

Regarding pre and post ringing, perhaps my wording wasn't as technically accurate as would be ideal. I am learning as I go here. I'm not saying that analog filters show precausal characteristics in the analog domain. Clearly we are talking about digital signal processing here. A better way to say that perhaps would be that linear phase low pass filters, in band limiting the signal, remove upper harmonics that can result in measurable pre and post ringing on sharp transients when played back. That is known as the gibbs phenomenon. I may not have that technically 100% correct and I'm not asserting that it would be audible or that the ringing would occur in normal transients in music (as opposed to square waves) or anything else. Just that it is something that happens with band limited signal processing. At least that much can be said. Well that's what our old friend Monty Montgomery says: https://wiki.xiph.org/Videos/Digital_Show_and_Tell#Bandlimitation_and_timing .

Again, none of the above is nearly sufficient for the types of assertions that Sergei has been making, as far as I can tell.