John Atkinson: Yes, MQA IS Elegant...

[rickca]

1 hour ago, james45974 said:

there had been waning interest

So what?  Waning interest in the birth of a new world.  People can be so obtuse.

[Miska]

In latest Stereophile J. Atkinson was once again going for the age old fallacy of looking only one aspect of the the filters (time domain) without putting the other aspect (frequency domain) side by side with it.

 

Just ignore anything that tries to discuss time domain without accompanying frequency domain plot.

[Miska]

Another aspect to make MQA look nicer was to use rolled off impulse to test the filters. Utilizing wide bandwidth microphones like Sanken CO-100K to record cymbals, soprano glockenspiel, metal claves, wood block and such from player's ear distance (50 cm to 1 m) is more representative. Assuming everybody listens classical music and thus that it's frequency characteristics would apply is also wrong. Lot of source material comes from software synthesizers and such that have never seen a single ADC. Or just something like heavy metal band. Reproducing these correctly is as important as reproducing classical music correctly.

 

For analyzing digital filter performance, I personally use high resolution analyzer with bandwidth in MHz range, 100 kHz is not sufficient for that purpose.

 

[John_Atkinson]

1 hour ago, Miska said:

In latest Stereophile J. Atkinson was once again going for the age old fallacy of looking only one aspect of the the filters (time domain) without putting the other aspect (frequency domain) side by side with it.

 

We have already examined the frequency-domain performance of sharp rolloff- filters and "leaky" ones like MQA's upsampling filter. As I wrote earlier in this thread, it would be helpful if people actually read what we have written instead of firing from the hip.

 

John Atkinson

Editor, Stereophile

 

[Miska]

1 minute ago, John_Atkinson said:

We have already examined the frequency-domain performance of sharp rolloff- filters and "leaky" ones like MQA's upsampling filter. As I wrote earlier in this thread, it would be helpful if people actually read what we have written instead of firing from the hip.

 

I've been reading all Stereophiles for I don't know how many years, so yes I've read it all.

 

But you really need to put the time- and frequency domain plots side by side if you like to compare things. And not expect that people will go back to old issues to see plots that don't actually match the same cases. Now you measured impulse domain behavior using different method, but you didn't publish frequency domain plots corresponding the same impulse responses.

 

So you should have computed frequency domain plots of all the impulse responses and plot those on the side of the time domain plot.

 

Also using a band limited test signal to test impulse response of filters is not very representative of the the filter performance in first place.

 

[crenca]

31 minutes ago, Miska said:

 

I've been reading all Stereophiles for I don't know how many years, so yes I've read it all.

 

But you really need to put the time- and frequency domain plots side by side if you like to compare things. And not expect that people will go back to old issues to see plots that don't actually match the same cases. Now you measured impulse domain behavior using different method, but you didn't publish frequency domain plots corresponding the same impulse responses.

 

So you should have computed frequency domain plots of all the impulse responses and plot those on the side of the time domain plot.

 

Also using a band limited test signal to test impulse response of filters is not very representative of the the filter performance in first place.

 

 

To expand upon this, why is the particular time/frequency trade off's of MQA any more "elegant" and faithful to the original signal (fidelitous) than the various choices other's make?  The man who admits that Audiophiledom is about "opposing belief" certainly has his it would appear.

 

What @John_Atkinsonhas explicitly admitted is that what he really finds "elegant" about MQA is how it was supposed to provide an "end to end" quality control of the whole recording/production/delivery/consumer play back chain.  JA is uncomfortable with the fidelity chaos that rules currently (as is every Audiophile) and thus is willing to sell his soul (so to speak - or has he in reality ? ?) to the closed source, proprietary, DRM, innovation strangling devil in order to clean up this chaos.  Tyranny in the name of progress.  Don't ever accuse JA of being a populist or a democrat (philosophy, not party).

 

What is head slapplying obvious though is that this devil does not actually do any of the things it says it does, so this "one ring to rule them all" is an empty promise (I won't go into in all the ways the mask has been pulled from the marketing speak).  So all JA has left is to peddle the "idea" of an MQA ring-of-power, because that actual execution is not just imperfect (as all things are), but a spectacular failure at actually bringing any order at all to this chaos. 

 

All of this has been known for a while now, so why are we still talking about this?  Thy do the trade publications still peddle it?  Ego fur sur, for how can JA back out of this deep deep hole he and Bob have dug?  There is that, but also the $investment$ that the trade publications are about as well...

[mansr]

36 minutes ago, John_Atkinson said:

if people actually read what we have written

Why don't you read what we have written for a change?

[daverich4]

On 8/15/2018 at 12:37 PM, Brinkman Ship said:

..except, there is no "if", virtually all non industry linked listeners feel not only does it NOT improve the sound, but it makes it worse.

 

Really? Virtually all? I don’t. 

[Brinkman Ship]

Way to go Archimago!

 

"Where's the "elegance" again in MQA?

 

Which part is the "brilliant innovation"?

 

How exactly is this lossy compression system "high resolution" even?

 

Presumably this is the last of Stereophile's series on MQA. Showing a few impulse responses "encapsulates" all there is to the claim of "deblurring", I suppose.

 

What else to say but time to move on."

 

https://www.stereophile.com/content/zen-art-ad-conversion#JlXqZWCETlhmggx7.99

[Brinkman Ship]

2 minutes ago, daverich4 said:

 

Really? Virtually all? I don’t. 

Of course you don't. Your posting history makes it clear you are a member of the MQA Defenders Of The Universe.

[daverich4]

2 minutes ago, Brinkman Ship said:

Of course you don't. Your posting history makes it clear you are a member of the MQA Defenders Of The Universe.

 

My posting history? What on earth are you talking about? 

[crenca]

14 hours ago, Brinkman Ship said:

Way to go Archimago!

 

"Where's the "elegance" again in MQA?

 

Which part is the "brilliant innovation"?

 

How exactly is this lossy compression system "high resolution" even?

 

Presumably this is the last of Stereophile's series on MQA. Showing a few impulse responses "encapsulates" all there is to the claim of "deblurring", I suppose.

 

What else to say but time to move on."

 

https://www.stereophile.com/content/zen-art-ad-conversion#JlXqZWCETlhmggx7.99

`

I guess his comment was deleted? 

[christopher3393]

18 minutes ago, crenca said:

`

I guess his comment was deleted? 

 

Just checked: it is still there.

[crenca]

1 minute ago, christopher3393 said:

 

Just checked: it is still there.

 

Your right, somehow I missed it...

[christopher3393]

6 minutes ago, crenca said:

 

Your right, somehow I missed it...

 

Could happen to anyone. My eyes often begin to glaze over when reading the comments sections of these articles.

For the most part: many words, little or no substance.

[Brinkman Ship]

Interesting exchanges between Doug S. of Soundstage and Atkinson-

 

https://www.audioasylum.com/cgi/vt.mpl?f=critics&m=93444

[Jud]

On 8/16/2018 at 4:38 PM, crenca said:

 

All of this has been known for a while now, so why are we still talking about this?

 

Considering the microscopic market share and mostly lackluster reaction, I do wonder about the level of attention given by the audio press and forum members here. 

[Brinkman Ship]

1 hour ago, Jud said:

 

Considering the microscopic market share and mostly lackluster reaction, I do wonder about the level of attention given by the audio press and forum members here. 

Chicken or egg? If the press never mentioned MQA again..don't you think MQA related posts would eventually sputter into zero?

[Rt66indierock]

1 hour ago, Jud said:

 

Considering the microscopic market share and mostly lackluster reaction, I do wonder about the level of attention given by the audio press and forum members here. 

 

Jud I want MQA Ltd liquidated. And I will continue to watch varying places to make sure MQA doesn't slipped in the backdoor as other DRM schemes have in the computer world. 

[Jud]

26 minutes ago, Brinkman Ship said:

Chicken or egg? If the press never mentioned MQA again..don't you think MQA related posts would eventually sputter into zero?

 

Or close to it, yeah. 

 

Just wondering why there's even the extent of reaction that occurs now when MQA is mentioned. I tried it,  thought it was slightly worse than the "real thing" except when the master was different.  At that point it became a big yawn for me, other than watching to see whether there was any chance it might achieve enough market power to crowd out other options.  So far that looks unlikely, and I'd be pleased if that's the way it turned out.

[crenca]

2 hours ago, Brinkman Ship said:

Interesting exchanges between Doug S. of Soundstage and Atkinson-

 

https://www.audioasylum.com/cgi/vt.mpl?f=critics&m=93444

 

Doug is right:

 

"...The big issue surrounded the MQA-audio-writer debacle is that right off the bat, for whatever reason(s), these writers seemingly bought in hook, line, and sinker to things like it being lossless (it's not), it being the size of a CD file (it's not -- it's just over 50% larger at 24/48), it contained data over 24/96 (all tests show it doesn't -- discarded on compression), and on and on. Charley, repeatedly, went on the record saying all these things weren't true, the filter was also nothing new, and that people had been duped.

Doug
SoundStage!"

 

John Atkinson in that thread - it's August 2018 folks! - is STILL asserting that MQA is "the only commercially available end-to-end solution".  Except, it is not really "end-to-end" at all, not in actual fact, only in marketing claims.  To add Doug's debunking, it turns out it is trivially easy to make the blue light "authentication" come on even when the file has been altered, end users DAC's are NOT being accounted for because only a handful of standard digital filters are being applied, and end users are still being lied to about the nature of what they are playing by being told MQA is somehow unfolding to >96 sample rate....

[John_Atkinson]

7 minutes ago, crenca said:

John Atkinson in that thread - it's August 2018 folks! - is STILL asserting that MQA is "the only commercially available end-to-end solution". 

 

Please note that time-domain performance was the context for this posting to the Audio Asylum, not DRM, not the possibility of aliasing, not the file size, not the lossy vs lossless argument. In that context, I wrote "MQA, _if_ it operates as describes and as I investigated in my article, is the only commercially available end-to end solution. (Unless you consider very high-bitrate DSD implemented with complementary first-order low-pass filters.)"

 

As you appear to be arguing with that assertion, what other combinations of commercially available A/D converter and D/A conversion, other than MQA, Ayre's experimental filters, or, possibly, high bit-rate DSD, give you perfect behavior in the time domain from analog original signal to the analog reconstruction of that signal?

 

John Atkinson

Editor, Stereophile

[FredericV]

25 minutes ago, John_Atkinson said:

As you appear to be arguing with that assertion, what other combinations of commercially available A/D converter and D/A conversion, other than MQA, Ayre's experimental filters, or, possibly, high bit-rate DSD, give you perfect behavior in the time domain from analog original signal to the analog reconstruction of that signal?

 

John Atkinson

Editor, Stereophile

Most speakers don't even get the time domain right. We have access to a speaker design which gets it right:

http://www.legendloudspeakers.com/

I was witness of the following:

- live drum recorded with a Zoom recorder
- played back via antelope audio DAC driven by foobar on a notebook
- playback on the active john watkinson legends, which get the time domain right

The drum sounded exactly like the real thing, with the same dynamics and transients. Not many speakers can do this. Elecstrostatic speaker can as they get the time domain right, but at a lower volume and they lack the bass extention.

MQA was not in the loop.

MQA is not needed for end to end

Fact is that speaker distortion is a far bigger issue than what MQA is trying to solve. Just look at the post-ringing of most tweeters, not being able to stop the motion in time after non-periodic sounds. Therefore messing up the time domain. Our hearing is very sensitive to non-periodic sound events as a survival mechanism, to locate the threat.

If you come to the X-FI show in The Netherlands, I will arrange for the co-designer of this speaker to repeat his demo.

The co-designer is also a teacher in the field of sound reproduction at the Brussels film school and a 3D stereography expert, and director of photography. His colleague wrote the bible on digital audio:

https://www.amazon.com/Digital-Audio-Third-John-Watkinson/dp/0240515870

 

Together they designed the John Watkinson Legends.

[Jud]

26 minutes ago, John_Atkinson said:

 

Please note that time-domain performance was the context for this posting to the Audio Asylum, not DRM, not the possibility of aliasing, not the file size, not the lossy vs lossless argument. In that context, I wrote "MQA, _if_ it operates as describes and as I investigated in my article, is the only commercially available end-to end solution. (Unless you consider very high-bitrate DSD implemented with complementary first-order low-pass filters.)"

 

As you appear to be arguing with that assertion, what other combinations of commercially available A/D converter and D/A conversion, other than MQA, Ayre's experimental filters, or, possibly, high bit-rate DSD, give you perfect behavior in the time domain from analog original signal to the analog reconstruction of that signal?

 

John Atkinson

Editor, Stereophile

 

John, first I'd quibble with "perfect." But let's suppose the filters ring little if at all.  Even if we say there are a group of listeners who are particularly sensitive to time domain distortions, that's not all we hear - any of us.  If the price of little or no ringing for these particular filters is imaging, aliasing, and higher THD, can we as people devoted to accuracy of reproduction ignore that?  (This is of course aside from any questions about what is being reproduced, since the MQA process alters the original file.)

[crenca]

39 minutes ago, John_Atkinson said:

As you appear to be arguing with that assertion, what other combinations of commercially available A/D converter and D/A conversion, other than MQA, Ayre's experimental filters, or, possibly, high bit-rate DSD, give you perfect behavior in the time domain from analog original signal to the analog reconstruction of that signal?

 

John Atkinson

Editor, Stereophile

 

I was speaking to that assertion, so I think you are correct...in a technical sense.  Your overall understanding around the "end-to-end" as you have argued here at CA and I believe in your more recent MQA ponderings (though I admit I tend to only read the snippets posted here and try to avoid giving Stereophile clicks) is more comprehensive.  This is why Doug and I responeded the way we did.

 

As far as "perfect behavior in the time domain", is it not not linear phase that actually interpolates (i.e. reconstructs) the band limited waveform in the "time domain" (putting aside arguments about the audibility of so called "ringing") and it is "minimum phase" that actually skews the frequencies so called "time domain" behavior? It is the "roll off" or "steepness" that affects both minimum and linear/accurate phase filtering algorithmic methods that is more relevant to the particular interpolating scheme that you and MQA are attempting to sell as I understand it.  All this is related from linear and non-linear distortion (which is BOTH "time" and "frequency" domain impactful) as well.  In other words, since there is no free lunch (real world signals are band/time limited) there is always a trade off between "time domain" and "frequency domain".  MQA is NOT a perfect deliverer of the original analog signal, though it makes choices trade-off choices like every other product/technique does. 

 

Not that I am really asking you (about the above of which I am probably in error about) as you and your writers are not reliable authorities on matters such as these.  It is obvious (as in the sky is blue obvious) that you and your writers (such as Jim A) fell for MQA first, and then have been trying to justify and sell it (technically, market position, etc.) after the fact.  This is the " MQA-audio-writer debacle" to which Doug S refers and for which you are STILL attempting damage control mode, unable to admit that you were roundly "duped" just as Doug points out.  This is in contrast to consumers and some (but not all) "industry insiders" as represented by this forum, who started asking real questions about MQA from the very beginning...