Article: MQA: A Review of controversies, concerns, and cautions

[crenca]

1 minute ago, miguelito said:

Perhaps it works is not really an endorsement is it?

 

I don't know, it is audiophiledom where something (even if it is just in your head) is always something!  Obviously not for me, but whatever it is doing will be justified in the minds of some as long as they hear a "vague explanation" as Archimago describes Bob S speak.  My point is that JA (or anyone else) can convince themselves by putting together a coherent, "rational" explanation from MQA charts, papers, etc. and then tell themselves and the world it does work.

 

 

[MikeyFresh]

7 hours ago, beetlemania said:

The mic preamp out is an analog signal as is the DAC output. But, in between, digital conversions are required. True fidelity seems a tall order, especially with a lossy codec. What am I missing?

 

You've missed nothing at all.

[WiWavelength]

1 hour ago, miguelito said:

...and they manage to sort out all of the information of each album (ADCs etc)...

 

I have my doubts about MQA's ability to research the recording provenance of albums slated for encoding.  Still, consider this an interesting aside.

 

Among my Stereophile CDs, I have the original "Test CD" [STPH-002-2] from 1990.  And on this CD is a 15 ips magnetic tape recording transferred to digital twice for comparison -- via the standard at the time Sony PCM-1630 ADC as well as via the then new Chesky 128x oversampling ADC.  Booklet text is credited to both John Atkinson (JA) and Robert Harley (formerly, RH), and in regards to the PCM-1630, the booklet quoth "...the initial A/D conversion for nine out of every ten CDs is made with this Sony converter."

 

I have no reason to doubt this assertion.  Per my understanding of early CD production, standard practice was to deliver to the pressing plant the digital master on U-matic tape, and the Sony PCM-1600 series provided the PCM adaptor for U-matic digital audio recording.

 

So, if a digital master is of a certain age, it likely had at least a final pass through a Sony PCM-1600 series ADC, perhaps even if upstream digital recording had been Soundstream, Mitsubishi, or 3M, etc.

 

AJ

[MikeyFresh]

1 hour ago, Archimago said:

LOL - perhaps it's precisely because they attended The Shard that they didn't question some of the obvious!

 

Yes, precisely.

[FredericV]

5 hours ago, miguelito said:

Consider that about 8,000 albums have been encoded in MQA. Imagine there are 10 teams doing this encoding, and they manage to sort out all of the information of each album (ADCs etc) such that they can encode 1 album a day per team. And these teams work 252 days per year - ie every business day each team cranks out an album. This means that it will take 3.1 years for these teams to do these 8,000 albums. Ponder...

The bulk of the albums is most likely batch converted, with some special cases where they will do more effort.

They need a story to sell, which is the provenance / white glove argument. So for a very limited amount of albums, they probably did more research into how the album was recorded / created, and have press like Hans Beekhuyzen write about it.

When the album is a complex mix of many AD/DA steps (such as effects processors on the effects bus of a mixing desk, instruments with sample banks such as most keyboards - basically everything that is non-acoustical or processed),  no way MQA is going to fix all those errors which were already downmixed into a 2 channel version.




 

[firedog]

5 hours ago, WiWavelength said:

 

I have my doubts about MQA's ability to research the recording provenance of albums slated for encoding.  Still, consider this an interesting aside.

 

Among my Stereophile CDs, I have the original "Test CD" [STPH-002-2] from 1990.  And on this CD is a 15 ips magnetic tape recording transferred to digital twice for comparison -- via the standard at the time Sony PCM-1630 ADC as well as via the then new Chesky 128x oversampling ADC.  Booklet text is credited to both John Atkinson (JA) and Robert Harley (formerly, RH), and in regards to the PCM-1630, the booklet quoth "...the initial A/D conversion for nine out of every ten CDs is made with this Sony converter."

 

I have no reason to doubt this assertion.  Per my understanding of early CD production, standard practice was to deliver to the pressing plant the digital master on U-matic tape, and the Sony PCM-1600 series provided the PCM adaptor for U-matic digital audio recording.

 

So, if a digital master is of a certain age, it likely had at least a final pass through a Sony PCM-1600 series ADC, perhaps even if upstream digital recording had been Soundstream, Mitsubishi, or 3M, etc.

 

AJ

MQA claim that the number of different ADCs used in old recordings was very small - as not many existed. So the issue of multiple unknown ADC's on old tracks isn't a problem.

 

They also claim they've analyzed so many tracks where they do know the ADC's used, that they've developed an algorithm that enables them to make a highly accurate guess about what ADC was used, even if provenance of the track isn't known. 

 

They can then apply the proper correction, and all of this is at least partially automated. 

[Fokus]

They claim a lot.

[rwdvis]

20 hours ago, John_Atkinson said:

But I fail to comprehend what the problem has been so far with our coverage of MQA....

 

John Atkinson

Editor, Stereophile

“It is difficult to get a man to understand something when his salary depends upon his not understanding it.”  -Upton Sinclair

 

[miguelito]

6 hours ago, WiWavelength said:

"...the initial A/D conversion for nine out of every ten CDs is made with this Sony converter."

 

I have no reason to doubt this assertion. 

I wouldn't doubt that either. Using a better ADC, especially in those days, would no doubt render a better result. However, the MQA argument is "we know the ADC so we can fix the artifacts it added to the digital stream". But once you've encoded with a less than optimal ADC, you've lost information. Yes, you might improve it a bit with some processing, but there's no substitution for remastering. I would be very skeptical MQA is able to correct anything in the newest ADC technology.

[miguelito]

1 hour ago, firedog said:

They also claim they've analyzed so many tracks where they do know the ADC's used, that they've developed an algorithm that enables them to make a highly accurate guess about what ADC was used, even if provenance of the track isn't known. 

 

They can then apply the proper correction, and all of this is at least partially automated. 

How is this not the same as DSP at the end of the chain?

[firedog]

1 hour ago, miguelito said:

How is this not the same as DSP at the end of the chain?

How is it the same?

They are claiming to correct the timing errors of the ADC in the recording.DSP at the end of the chain has a different purpose.

[miguelito]

36 minutes ago, firedog said:

How is it the same?

They are claiming to correct the timing errors of the ADC in the recording.DSP at the end of the chain has a different purpose.

Consider an ideal (maybe realistic for an analog recording):

1- mic--> analog mixing console --> ADC

2- PCM stream delivered as a file to you

3- Playback

 

And consider the process in MQA:

a- mic--> analog mixing console --> ADC

b- PCM --> deblurred PCM

c- MQA lossy compression

d- MQA file delivered to you

e- Playback over MQA decoder

 

Why can't I put step 'b' in step '3'? Are you saying that the lossy MQA compression is inextricable from 'b'? Why would that be? And if there's some processing done in 'c' that is part of 'b', why couldn't I still put it in '3'?

 

Explain. Please don't refer back to Bob Stuart BS (pun intended).

[miguelito]

8 hours ago, MikeyFresh said:

10 hours ago, Archimago said:

LOL - perhaps it's precisely because they attended The Shard that they didn't question some of the obvious!

 

Yes, precisely.

I think it is also fair to say that what is presented as "A/B" examples in such gatherings is suspect. When I listened to "A/B" comparisons at Meridian NYC there was no contest - the MQA version were markedly better. After a lot of listening on TIDAL (and full hardware decoding) I can't reproduce those differences. So my conclusion is they were showing cases where the files were not only MQA encoded but fully remastered.

[Fokus]

1 hour ago, miguelito said:

Consider an ideal (maybe realistic for an analog recording):

1- mic--> analog mixing console --> ADC

2- PCM stream delivered as a file to you

3- Playback

 

And consider the process in MQA:

a- mic--> analog mixing console --> ADC

b- PCM --> deblurred PCM

c- MQA lossy compression

d- MQA file delivered to you

e- Playback over MQA decoder

 

Why can't I put step 'b' in step '3'?

 

Because the 'deblurring' requires an input at sample rate higher than 96kHz. Which is something you don't get your hands on at stage 3 under the new order.

 

As for any (real or imagined) corrections done to originals of 96kHz and lower: yes, all of these can be implemented anywhere in the flow. These are conceptually not different from any mastering step. For instance, if MQA have a magic plugin that makes any PCM1630 recording better, then they could sell that plugin to each and every mastering studio. Only ... MQA are not interested in this business model.

 

[crenca]

1 hour ago, miguelito said:

Consider an ideal (maybe realistic for an analog recording):

1- mic--> analog mixing console --> ADC

2- PCM stream delivered as a file to you

3- Playback

 

And consider the process in MQA:

a- mic--> analog mixing console --> ADC

b- PCM --> deblurred PCM

c- MQA lossy compression

d- MQA file delivered to you

e- Playback over MQA decoder

 

Why can't I put step 'b' in step '3'? Are you saying that the lossy MQA compression is inextricable from 'b'? Why would that be? And if there's some processing done in 'c' that is part of 'b', why couldn't I still put it in '3'?

 

Explain. Please don't refer back to Bob Stuart BS (pun intended).

 

It was @Doug Schneider(going from memory) who quoted Bob S as saying "we are not going to sell software tools".  In other words, Bob S wants to sell an "end to end" DRM format.  So there is your answer.

 

edit:  I should add that just as Bob S compromised on MQA around hardware only unfolding, business realities might just force a compromise on this as well.  On the other hand,  if there is not really anything to it (other than already known art), separating the compression from the phase correction would do them no $good$.

 

This just might be the genius in JI's (and @John_Atkinson??) strategery of a demand for a separation - force MQA's hand to call their bluff that "deblurring" is a couple of low cards...

 

[miguelito]

30 minutes ago, Fokus said:

Because the 'deblurring' requires an input at sample rate higher than 96kHz. Which is something you don't get your hands on at stage 3 under the new order.

Where did you get that from?? So you're saying that deblurring cannot be done on any recording done to redbook format such as Dire Straights's "Brothers in Arms"??? It is precisely in these early digital cases where it would make the most sense.

 

30 minutes ago, Fokus said:

As for any (real or imagined) corrections done to originals of 96kHz and lower: yes, all of these can be implemented anywhere in the flow. These are conceptually not different from any mastering step. For instance, if MQA have a magic plugin that makes any PCM1630 recording better, then they could sell that plugin to each and every mastering studio. Only ... MQA are not interested in this business model.

"Magic" is not in my vocabulary. Maybe you mean "eq"? 

[Archimago]

55 minutes ago, Fokus said:

... For instance, if MQA have a magic plugin that makes any PCM1630 recording better, then they could sell that plugin to each and every mastering studio. Only ... MQA are not interested in this business model.

 

And that is exactly what this should all have been about. A plugin used by the artist and studio in the production. "Authentically" chosen to be used under their auspices during a project and mixed into the recording as a conscious choice.

 

Not some kind of after-the-fact DSP processing that might or might not be of benefit and inextricably bundled in with the compression system.

 

Their current business model has way more tentacles for extraction of potential revenue than just selling a plugin ...

 

 

[Archimago]

42 minutes ago, miguelito said:

Where did you get that from?? So you're saying that deblurring cannot be done on any recording done to redbook format such as Dire Straights's "Brothers in Arms"??? It is precisely in these early digital cases where it would make the most sense.

 

"Magic" is not in my vocabulary. Maybe you mean "eq"? 

 

But that depends on what "deblurring" is, right? Which we don't have a clear definition of from MQA. Without knowledge of the mechanism by which they're "correcting" the sound, we cannot really predict the domain of the kinds of recordings this could impact nor the magnitude of said potential improvements.

 

Remember that although they call it "deblurring" and show a bunch of impulse responses claiming the importance of the time-domain component, we could be looking at a software suite with things like pitch correction (referred to with the Toneff & Dobrogosz "Fairytales" article in Stereophile), subtle EQ, as well as application of group delay changes. 

[crenca]

1 minute ago, Archimago said:

 

But that depends on what "deblurring" is, right? Which we don't have a clear definition of from MQA. Without knowledge of the mechanism by which they're "correcting" the sound, we cannot really predict the domain of the kinds of recordings this could impact nor the magnitude of said potential improvements.

 

 

 

Would you (or anyone else) say that besides the above, there could be a real reason for it being an encoding process only?  It seems to me that is only possible if it is truly something new in signal processing.  Everything that is realistically could be can also be done in software tools at playback time...

[Doug Schneider]

39 minutes ago, crenca said:

 

It was @Doug Schneider(going from memory) who quoted Bob S as saying "we are not going to sell software tools".  In other words, Bob S wants to sell an "end to end" DRM format.  So there is your answer.

 

Correct. Upwards in this thread somewhere. It came from when I talked to Bob Stuart on the phone, just prior to publishing my first MQA article, in April 2016. We talked about the compression and the time-based construction being separate things that could be done independently. It was at that point that he told me they weren't interested in selling tools, meaning they wanted the entire thing to be considered one.

 

Doug

[Archimago]

25 minutes ago, crenca said:

 

 

Would you (or anyone else) say that besides the above, there could be a real reason for it being an encoding process only?  It seems to me that is only possible if it is truly something new in signal processing.  Everything that is realistically could be can also be done in software tools at playback time...

 

Agree.

 

This was all software from the start. Other than those "white glove" recordings that were specifically remixed, if these are just after-the-fact adjustments, no reason why we can't take the 24-bit "studio master" and apply some kind of DSP ourselves with appropriate parameters to reconstruct whatever MQA feels "deblurring" is - whether it be actual time-domain, subtle frequency shift, certain type of noise shaped dithering, certain upsampling digital filtering, etc...

 

They'd rather preserve the mystique and sell the gnosticism .

[Dr Tone]

2 hours ago, miguelito said:

Why can't I put step 'b' in step '3'? Are you saying that the lossy MQA compression is inextricable from 'b'? Why would that be? And if there's some processing done in 'c' that is part of 'b', why couldn't I still put it in '3'?

 

I doubt there is a reason they can't.  They reason they want it done up front is so they can lock it down and license it.  It's all about the Benjamins.

[miguelito]

1 hour ago, Archimago said:

we could be looking at a software suite with things like pitch correction (referred to with the Toneff & Dobrogosz "Fairytales" article in Stereophile), subtle EQ, as well as application of group delay changes. 

This is my impression after listening to a lot of solo piano on MQA vs their non-MQA counterparts - similarly to oversaturating a picture. One notable example is Keith Jarrett's "Koln Concert" - the 24/96 remaster (HDTracks) sounds a bit "dryer" than the MQA version. I get the impression that the dryer version is more realistic (especially given the history of this album - Jarrett played on an inferior, mistuned piano with some non-working keys).

 

https://en.wikipedia.org/wiki/The_Köln_Concert

[miguelito]

1 hour ago, Archimago said:

Remember that although they call it "deblurring" and show a bunch of impulse responses claiming the importance of the time-domain component, we could be looking at a software suite with things like pitch correction (referred to with the Toneff & Dobrogosz "Fairytales" article in Stereophile), subtle EQ, as well as application of group delay changes. 

I read this review on the magazine. All I can say is "Bravo!" for the restoration of this recording. I will also say that's there's nothing that standard PCM could not do here - I much rather get the full-res restored version than the MQA version.

 

Sadly, all I can find is an AAC version of the MQA release on iTunes! So not only is it 16/48, it is 16/48 already compressed! 

 

Take away that Bravo!!!

[adamdea]

4 hours ago, Fokus said:

 

Because the 'deblurring' requires an input at sample rate higher than 96kHz. Which is something you don't get your hands on at stage 3 under the new order.

 

As for any (real or imagined) corrections done to originals of 96kHz and lower: yes, all of these can be implemented anywhere in the flow. These are conceptually not different from any mastering step. For instance, if MQA have a magic plugin that makes any PCM1630 recording better, then they could sell that plugin to each and every mastering studio. Only ... MQA are not interested in this business model.

 

What sort of AA filter did the pcm 1630 have?