MQA is Vaporware

[KeenObserver]

6 hours ago, Lee Scoggins said:

 

Paul, I talked to two different streaming services executives and they told me that the size of the file does matter and they do have issues at scale.  So it's hard for me to take your word for it based on this research.

 

As for what the MQA team should do, they did share a lot of information with Chris from Day 1 including the lengthy Q&A from Bob Stuart.  But they feel there is not a fair debate to be had on this forum so they are ignoring it.

 

Translated from Lee speak to normal speak: MQA does not care to have AS expose the truth about MQA.

 

MQA prefers to have forums repeat the buzzwords and talking points that MQA provides.

They prefer blogs that put out MQA press releases as editorial content.

They prefer forums that are incapable of critical thinking.

[Miska]

6 hours ago, manisandher said:

 

That's great. Thanks Paul.

 

So, it seems I didn't do too bad a job at aligning the files manually:

 

The Delta Phase plot looks interesting:

 

Anyone have any thoughts about this?

 

Mani.

 

Been through some kind of minimum phase filter?

[KeenObserver]

4 hours ago, Lee Scoggins said:

 

They usually release white papers at AES meetings.

 

Scott paper that gets flushed down the MQA "ecosystem".

[mansr]

1 hour ago, mcgillroy said:

Not sure if this was discussed here already but a new patent application by Peter Craven and Bob Stuart was published late February:

 

DIGITAL ENCAPSULATION OF AUDIO SIGNALS - http://www.freepatentsonline.com/y2019/0057709.html

 

See also here: https://www.gearslutz.com/board/showpost.php?p=13859201&postcount=541

From a brief look, that looks like a description of the "rendering" filters used in MQA.

[Miska]

5 hours ago, mansr said:

No more than standard FLAC with a little (shaped) dither can provide. Moreover, a 48/24 source file actually becomes bigger after MQA coding, and these account for a large fraction of the catalogue.

 

Just run again one of my favorite test tracks through.

 

Original FLAC just TPDF dithered to 18-bit, still 96/24 FLAC container; size 72 MB.

MQA FLAC; size 70 MB, with the "mqascan" reporting "MQA signature at bit 9" (and just for fun, render filter is 4).

 

I'd say the 18-bit version of the original is much better quality/resolution than the MQA, with practically same size.

 

[KeenObserver]

4 hours ago, Lee Scoggins said:

 

No.  That would be difficult to get evidence for since it is a change in behavior in the future.  But the label people I have spoken to recognize that paying the artists more is required for the system to be healthy.

 

Thank God I left my hip boots on!

[Axial]

Bummer, they need a new adaptor/dongle ...

• https://darko.audio/2019/03/the-inconvenient-truth-about-mqa-on-ios/

[opus101]

6 hours ago, Lee Scoggins said:

 

But the label people I have spoken to recognize that paying the artists more is required for the system to be healthy.

 

Healthy being less profitable than unhealthy it's a reasonable assumption they're content to stick with the latter.

[Fokus]

11 hours ago, Lee Scoggins said:

 

They usually release white papers at AES meetings.

 

MQA doling out white papers at AES? LOL!

 

Apparently you don't know the difference between a paper and a white paper.

 

[Ralf11]

maybe he meant toilet paper

[manisandher]

7 hours ago, Miska said:

Been through some kind of minimum phase filter?

 

OK, if so, nothing that we couldn't do ourselves during playback.

 

Whatever's causing what's seen on the delta-phase plot, its affect seems to be audible - to my ears, the MQA sounds different to the hires. Actually, everyone who took the blind test reported hearing differences, and the descriptions were pretty consistent.

 

I think those who consider MQA a simple compression scheme are mistaken. It seems to be doing more...

 

Mani.

[Miska]

1 hour ago, manisandher said:

OK, if so, nothing that we couldn't do ourselves during playback.

 

Whatever's causing what's seen on the delta-phase plot, its affect seems to be audible - to my ears, the MQA sounds different to the hires. Actually, everyone who took the blind test reported hearing differences, and the descriptions were pretty consistent.

 

I think those who consider MQA a simple compression scheme are mistaken. It seems to be doing more...

 

Of course, given the original hires, we could do all the same at playback time, but not sure if I'd want to do all of that...   (like aliasing)

 

The difference where you draw a line between what is part of compression scheme and what is part of "more" is a bit "blurry" too. Because of the folding stuff... When the source is higher than 88.2/96k rate they also need to decimate it down to the 88.2/96k rate before going for the encoding stage. Once there they need to do a band split to fold the upper band down and as side effect they have also fair bit of aliasing. And then the encoder needs to decide how many bits it needs to shave off the 24-bit base-band to encode and store the upper band information. More there is to encode, more it needs to limit the resolution. However, this is partially helped by the earlier filter (probably this minimum phase) that rolls off during the upper band, so amount of content from there is reduced already.

 

Most of this quality degradation is done practically to support the folding-compression scheme...

 

Just like in MP3 or AAC, lot of things are "done" to support the compression. For example MP3 encoders typically apply 16 kHz low-pass filter when encoding for bitrates of 128kbps and below. This is because encoding those highest frequencies takes proportionally the most bandwidth, while hearing is not the most sensitive there. Generally most MP3 and AAC faults can be heard on highs and transients. But you could implement this 16k filter also in many different ways, for example either using linear or minimum-phase filter!

 

[mansr]

3 hours ago, manisandher said:

I think those who consider MQA a simple compression scheme are mistaken. It seems to be doing more...

It is certainly more than a simple compression scheme. We know it does these things:

1. Downsample the source to 88.2/96 kHz using leaky filters, likely minimum phase.
2. Split into high and low bands.
3. Compress high band into bits 0-7 (typically).
4. Apply pseudo-random dither to mask compressed data as noise.

5. Apply peak limiting to low band, store compressed correction data in bit 8 (typically).

6. Encode various parameters needed by decoder into bit 8.

7. Compute authentication signatures, store in bit 8.

The decoder does the reverse of these steps. At any point in this process, the encoder might also do additional things that don't have a matching stage in the decoder. This is likely where the supposed ADC correction happens. If it exists, that is. Some MQA files appear to have gone through a denoising process. It is unclear whether this is considered part of MQA or just something they did anyway.

[crenca]

Thinking about starting an MQA "now what" thread.  After reading the two excellent posts above by Miska and Manser, or Jud's two "summaries" earlier, it is apparent we have matured past the technical exploration and refutation stage.  At this point, anybody who maintains that there is a technical justification for any of MQA's aspects is either ignorant or a confidence man.  

 

Now that we know MQA is bad for the consumer and bad for most in the industry itself (such as software/hardware providers), what now?  Sure, threads such as this will continue to educate the ignorant, but they are not going to change the minds of confidence me such as Lee or folks such as Quint and JA because they are too deep in the industry and audiophile "community" myths to do anything other than root for the MQA team over and against the consumer.  

 

What would a mature discussion past the technical stage look like?  What can (if anything) consumers do now?  Endless "debate" with the Lee's, Quint's, and JA's of this world is certainly not it - that is the very definition of futility.  What is the way forward?  

 

What say you?

[The Computer Audiophile]

17 minutes ago, crenca said:

Thinking about starting an MQA "now what" thread.  After reading the two excellent posts above by Miska and Manser, or Jud's two "summaries" earlier, it is apparent we have matured past the technical exploration and refutation stage.  At this point, anybody who maintains that there is a technical justification for any of MQA's aspects is either ignorant or a confidence man.  

 

Now that we know MQA is bad for the consumer and bad for most in the industry itself (such as software/hardware providers), what now?  Sure, threads such as this will continue to educate the ignorant, but they are not going to change the minds of confidence me such as Lee or folks such as Quint and JA because they are too deep in the industry and audiophile "community" myths to do anything other than root for the MQA team over and against the consumer.  

 

What would a mature discussion past the technical stage look like?  What can (if anything) consumers do now?  Endless "debate" with the Lee's, Quint's, and JA's of this world is certainly not it - that is the very definition of futility.  What is the way forward?  

 

What say you?

Was thinking of something similar myself. A thread people can digest if they just want to see MQA for what it is. 

[Samuel T Cogley]

18 minutes ago, crenca said:

What would a mature discussion past the technical stage look like?  What can (if anything) consumers do now?  Endless "debate" with the Lee's, Quint's, and JA's of this world is certainly not it - that is the very definition of futility.  What is the way forward?

 

It seems Scoggins is now hedging on his "predictions" of MQA's future success.  He's the most enthusiastic supporter of MQA that I'm aware of outside of MQA's executive staff, so this is news in my opinion.

 

Perhaps he's still feeling the sting of Light Harmonic's death spiral in spite of his fawning article.  And maybe he's seeing an outcome where all his MQA cheerleading comes back to haunt him as well.

[crenca]

3 minutes ago, The Computer Audiophile said:

Was thinking of something similar myself. A thread people can digest if they just want to see MQA for what it is. 

 

If by "what it is" you mean a technical explanation along the lines of the two posts above (and Jud's - Jud's actually uses more understandable layman language) I think that would be helpful - the 4 posts could serve as the OP.  A non "debating" thread where only technical explanations by qualified individuals are posted, perhaps actual tech questions allowed by non astroturfers.  

 

I was actually thinking of something more, a thread where we discuss "what now"  - what, if anything, can consumers do now given the technical stuff is at a mature point.

[The Computer Audiophile]

15 minutes ago, crenca said:

 

If by "what it is" you mean a technical explanation along the lines of the two posts above (and Jud's - Jud's actually uses more understandable layman language) I think that would be helpful - the 4 posts could serve as the OP.  A non "debating" thread where only technical explanations by qualified individuals are posted, perhaps actual tech questions allowed by non astroturfers.  

 

I was actually thinking of something more, a thread where we discuss "what now"  - what, if anything, can consumers do now given the technical stuff is at a mature point.

Perhaps starting out with the easily digestible Jud explanation leading into "what now..."

 

The facts are a good foundation for the what now discussion. Without them we'd all be drinking the Kool-aid. 

[Samuel T Cogley]

Just now, The Computer Audiophile said:

Perhaps starting out with the easily digestible Jud explanation leading into "what now..."

 

And @Jud deserves a BIG shout out for putting that together! 👍

[The Computer Audiophile]

Just now, Samuel T Cogley said:

 

And @Jud deserves a BIG shout out for putting that together! 👍

Absolutely!

[Paul R]

49 minutes ago, The Computer Audiophile said:

Was thinking of something similar myself. A thread people can digest if they just want to see MQA for what it is. 

 

I think it would be a good idea, but it needs to be moderated fairly heavily.

 

Just saying that we don't need all the personal attacks and abrasiveness, calling people ignorant, and finger pointing. Not if it is intended to be a place to attract people for real information. 

 

 

-Paul

 

[MikeyFresh]

5 minutes ago, Paul R said:

Just saying that we don't need all the personal attacks and abrasiveness, calling people ignorant

 

Not only has that always been the case via the site's rules, but I'm at a loss to identify this supposed preponderance of attacks and the like. Who was called ignorant and when?

 

I must have missed something.

[Rt66indierock]

Just now, MikeyFresh said:

 

Not only has that always been the case via the site's rules, but I'm at a loss to identify this supposed preponderance of attacks and the like. Who was called ignorant and when?

 

Well Paul R attacked me right of the audio journalists handbook.

[Paul R]

1 hour ago, crenca said:

 At this point, anybody who maintains that there is a technical justification for any of MQA's aspects is either ignorant or a confidence man.  

 

 

2 minutes ago, MikeyFresh said:

 

Not only has that always been the case via the site's rules, but I'm at a loss to identify this supposed preponderance of attacks and the like. Who was called ignorant and when?

 

I must have missed something.

 

You certainly did. 

 

[Paul R]

4 minutes ago, Rt66indierock said:

 

Well Paul R attacked me right of the audio journalists handbook.

 

And the farmer hauled yet another load away.

 

This is the kind of patented nonsense that is giving this place a bad reputation. 

 

Just saying, we don't need it. Not if you want a thread that may actually influence people to at least think about the facts and draw sane judgements. 

 

-Paul 

 

And, that is all I will say on the subject.  People do have the right to go to hell in whatever manner they choose. I am just sad to see this wonderful place dragged down with it.