John Atkinson: Yes, MQA IS Elegant...

[adamdea]

22 hours ago, John_Atkinson said:

 

If you read my article (which isn't specifically about MQA but about A/D conversion in general), you should note that the filter I mention that is free from ringing is

...

t if removing so-called "temporal bur" is indeed something that improves sound quality for the reasons I explain in the article,  the MQA process is one of the few commercially available end-to end solutions that would do that. If...

 

John Atkinson

Editor, Stereophile

 

I have re-and re -read tyour article John and I still can;t see the point. Everyone knows you can have a filter (or non-filter) with no temporal blur (if that exists) and that is by having a sample and hold circuit or some other form of weak or non-filter. So what?

I do think it's a good idea to try to use a transient that is not a dirac to model time domain behaviour , but I remain somewhat confused as to why you felt that the ones you used may have approximated to any real transient. I'm not sure that just having a downward sloping frequency/ amplitude plot is enough. Of course real transients don't come out of absolute silence. But if you can't show ringing with a real transient then maybe that's a clue.

[adamdea]

On 8/23/2018 at 8:34 PM, John_Atkinson said:

 

You can see from the article of mine that triggered this thread an example where a perfectly legal, band-limited impulse nevertheless excites the DAC reconstruction filter's sinc-function ringing.

 

John Atkinson

Editor, Stereophile

 

I think I read your article and can’t recall a band limited impulse. Can you refer to the precise text and/or figure number. 

I don’t have the article to hand but I thought from memory it had a downward sloping frequency response but did not have an absolute frequency limit. 

Apologies if I have misremembered.

[adamdea]

5 hours ago, John_Atkinson said:

 

Fig.12 at https://www.stereophile.com/content/zen-art-ad-conversion-page-2

 

John Atkinson

Editor, Stereophile

 

Sorry - but that seems to me to be the "impulse response" of a dac fed *something*. The something seems to be a non band limited impulse fed to a QA at 96khz and then sample rate converted to 44.1. Why are you calling this a perfectly legal band limited impulse" ?-it simply seems to be a conflation of the impulse responses of the decimation filter fed an illegal impulse and then whatever the dac does with it . Surely the ringing is the result of feeding a non band-limited impulse (whose frequency response is at figure 5 - it seems very clearly not to be band limited)  into a sample rate converter which presumably applies a brick wall 44.1khz filter? You have not isolated that ringing (ie the ringing  implicit in the 44.1khz data) from any ringing caused by the dac.  

In any event the prior question (which i think I asked above, but perhaps forgot) is why the original "impulse" (whose frequency response is at figure 5) is supposed to represent a musical transient. 

You say "With a musical transient, ie, when there is silence then data". Is a musical transient *silence* followed by data- where does that happen (even supposing that the rest of the orchestra stops playing before the transient), in space? 

It would be useful to see all of this demonstrated with a real musical transient

[adamdea]

38 minutes ago, John_Atkinson said:

 

Fig.12 in my article plots the sample values in the digital domain. No DAC involved.

 

 

To create those digital-domain data, I digitized at 96kHz a unidirectional, shaped analog pulse with an approximate bandwidth of 60kHz with the Ayre QA9 A/D converter with its "Listen" antialiasing filter. I then sample-rate-converted those data to 44.1kHz. This is described in the 

 

Apologies about the dac - but the confusion arose from

your original quote which I still can’t understand. But the ringing in figure 12 is just the src faced with an illegal impulse it seems 

[adamdea]

13 hours ago, John_Atkinson said:

 

The inference to be drawn is that every musical transient in a CD master will be accompanied by sinc-function ringing at Nyquist, either from the original A/D converter's anti-aliasing filter (if the recording was made at 44.1kHz), or from the sample-rate converter's low-pass filter used to create the master from 2Fs or 4Fs files. It seems incontrovertible, therefore that that ringing will excite the playback DAC's reconstruction filter, which will impose its own ringing on musical transients.

Read more at https://www.stereophile.com/content/zen-art-ad-conversion

 

John Atkinson

Editor, Stereophile

 

That part of your article is not clearly illustrated or argued. You do not clearly distinguish between ringing in the 16/44 data output of the src and that imposed by the dac (other than by passing assertion).

 

I seem to remember that linear phase filters are idempotent? I thought that meant that a second linear phase filter does not need to impose any ringing.

 

As fokus points out, the sampling theorem requires the signal to have no energy *at* Nyquist. If the signal has no energy at nyquist and the dac’s filter reaches stopband at nyquist, why will the dac’s filter ring? 

 

Conversely the point which your article seems to skate over is that if there is ringing in the 16/44 data then a slow/lazy/non  filter in the dac will just pass the ringing and there is no time domain advantage in such a filter. So the time domain blather about such filters is pure bull. 

 

And who says your made up shape is a musical transient? Where is an example of a genuine musical transient with that shape?  If every musical transient is accompanied by sinc- function ringing,  then all you would need to do to show this would be to show the time plot of *any* musical transient. If not then “the inference [is not] to be drawn”.

 

[adamdea]

14 minutes ago, John_Atkinson said:

 

 

The experimental evidence I presented is incontrovertible. That unless the user of an A/D converter is prepared to accept the possibility of some aliased image energy in order to use an antialiasing filter that preserves the time-domain behavior of the original analog signal, the resultant digital data will have sinc-function content at the Nyquist frequency accompanying every musical transient.

 

I can’t see how this could be described as  incontrovertible- it seems to me doesn’t even get off the ground because you did not start with a musical transient, let alone one captured with a real microphone.

 

Or does this only appear to me because of my (undoubted) lack of knowledge? 

[adamdea]

32 minutes ago, John_Atkinson said:

 

That is correct. I used an artificially generated, non-musical test signal with the necessary properties to investigate the subject in a repeatable and diagnostic manner. The use of such signals to investigate the behavior of audio components and infer the results of that behavior with music is routine. You can find myriad examples in the review archive at www.stereophile.com.

 

John Atkinson

Editor, Stereophile

 

Of course. But the claim that the results led to an incontrovertible conclusion about every actual musical event of a particular type is novel. You know this, why play act?

[adamdea]

2 hours ago, John_Atkinson said:

In the context of testing audio components, this is an enormous subject. All I can suggest is that you read the tutorial articles in the free on-line archives at the Stereophile website.

 

John Atkinson

Editor, Stereophile

 

This is a an obfuscation- you have explicitly  claimed that the results your test signal incontrovertibly led to a specific conclusion about musical signals. 

You have quite obviously overreached yourself in this claim. It might be that a more moderately worded claim would hold some water, and the smart thing to do might be to row back slightly.

[adamdea]

22 hours ago, Fokus said:

 

The synthetic nature of the signal is not the problem. As said before, one could take this signal and write a symphony for it.

 

The issue is 1) a far too complex approach for proving something very simple, and 2)  the employment of an improper SRC or failing to note a key property of that SRC and how this invalidates the 'proof'.

 

I get your 1) and 2). But the conclusion is couched in terms of “every musical transient” and I still cannot see how the generalisation could be “incontrovertible” even if the test could be described as *a* musical transient. My issue is not so much whether it is synthetic but whether it is representative. 

 

It is also worth noting a point you made earlier that the analysis does not include a transient band limited using a minimum phase filter in the src. If it had done so, any pre-ringing from the reconstruction filter could have more clearly been identified. And assuming that the data were properly band limited below nyquist, we could see whether it is indeed necessary to have some pre-ringing following reconstruction. 

[adamdea]

7 hours ago, crenca said:

 

Would you (or anyone else) say that this is the crux of the matter, to what extent typical low level ringing (putting aside the fact that good practices in delivery chain make it unnecessary) is "audible" in a an actual synthetic signal?

I think there are a number of important questions about pre-ringing in the 20ish kHz range (perhaps 18khz upwards)

1) is it audible at all for anyone who matters. Duration? Forward masking ? 

2) to what extent it really even arise in real world recordings

3) to the extent that it does arise, even if we imagined it to be audible, could it possibly account for claimed benefits of recordings claimed not to have it.

 

I see the expression “musical transient” with extreme suspicion. Part of the brouhaha about non orthodox filters is avague idea about smearing and timing and whatever. The insinuation is that 16/44 and or digital generally doesn’t have that toe tapping flat earth analogue musicality blah blah. And if the propaganda mood music is to the effect that the rhythm of music, the drum beats, the riff etc is all being spoiled. Hence the “musical transients”. But just look at any time domain representation of a track on audacity and try to find it. 

How often do you see a silence followed by a full scale transient?  Perhaps at the beginning of the fanfare for the common man, or in a Steve reich piece with wood blocks hitting together. Most of the time there’s some sound going on before the next note is played. Does it occur often enough or enough recordings to match the phenemenon “digital sounds bad” which it was supposed to explain. 

And once you put in your meridian style apodising filter, does it all go away? Does it happen at all in digitised versions of old rescordings. Even on new recordings, is there pre -ringing on every musical transient? Or does musical transient mean something which didn’t actually happen much in music. Even if pre-ringing were audible at 20 kHz what would this really explain? 

[adamdea]

2 hours ago, Ralf11 said:

What are the sharpest attacks ??

 

And what is the rise time for the sharpest part?

It is my understanding that castanets are the nearest you will ever see to an impulse. I don’t listen to a lot of castanet music. 

[adamdea]

3 hours ago, mansr said:

All very good questions. The answers are mainly "no."

As someone once said of the hymn “Jerusalem”.

[adamdea]

4 hours ago, Miska said:

 

This is article discussing the topic:

https://www.cco.caltech.edu/~boyk/spectra/spectra.htm

 

(the most valid for this is the section VI. Instruments without harmonics

https://www.cco.caltech.edu/~boyk/spectra/11.htm#b

 

Miska 

thanks for reminding us of that article. There are a couple of points which occur to me. 

- if we look for example at the jazz rimshot shown at figure 12 b) , am I reading this correctly: it appears to show a rise time of 148 us ie 6 and a bit sampling intervals at 44.1 kHz. That gives me the impression that the overall envelope could be captured in 44.1 kHz without too much problem. Certainly the envelope does not look like anything like a pulse. I guess there may still be enough of a discontinuity but how much ringing would that generate? 

Also at what level would it me reasonable to regard ringing as insignificant- could you resolve a tone at below ambient noise level if it only lasted for say 3 ms? 

 

I did play around a while ago with trying to create preringing in audacity by downsampling tracks but in the end when I proudly presented my homework to Jim lesurf he politely but firmly told me that I shouldn’t worry my pretty little head about it and that even the thing which looked like pre-ringing to me could not be shown to be. I can’t now remember what the reason was but I think it was to do with the very short duration and the potential for windowing artefacts etc. 

 

Another question is what percentage of sound over 20kHz (table 1) is enough to matter. The piano for example is shown to have some high frequency components it’s output. But with 0.02% over 20kHz it does make one wonder whether it really matters. Certainly the piano recordings I have looked at seem to have no sigificant stuff over 20 kHz. 

 

[adamdea]

3 hours ago, mansr said:

Wrong question. Content in the filter stopband does not cause "ringing." What matters is the intensity at the filter cut-off.

Yes I get that.

For most instruments the intensity will tend to decline with frequency though,  and one might be forgiven for the sin of lazily assuming that the proportion of energy over 20kHz will be a rough indicator of the intensity of energy at 22.05. But I can see that this may not hold true. That said It does establish a ceiling for the about of energy at nyquist