MQA: A Review of controversies, concerns, and cautions

[firedog]

You don't seem to have any use for facts. In format making, prior experience in the field is everything. To the extent people want to make market/business statements as Archimago made, and folks in this thread keep making, that is super relevant. You can't dismiss it and equal it to lay arguments you are making. This is the real world, not a made up monopoly game.

Bob has shown his skills in getting all these design wins from labels to music distributor to many devices. That is a fact and is based on the factors I mentioned.

Some of you hate to hear that Bob is exceptionally qualified and respected signal processing expert. You rather him be a no-name guy so that they can throw rocks at him easily. Watch Paul's video I post earlier in how people in the industry don't dare doing that:

"Bob stuart ... some of extremely bright people. One of the most knowledgeable bright people I know in our industry. Super nice guy. I really like Bob a lot. He has done a lot of terrific things. We as a company support MQA because our customers asked for it."

His prior work in standardization of formats he developed opens doors for him that would not be open to others. His deep knowledge of signal processing and psychoacoustics overcomes internal objections from people who know less than him.

So don't waste your breath on me with these debating tactics. You are not being told to trust MQA. You are being told to learn who you are dealing with and the powers in play in the industry. He got Paul to support MQA even though he says it is no good.

Paul only allowed the “first unfold” in his streaming module. Nothing else. And is on public record with negative comments about MQA.

Bob, Bob, Bob......so what. His commercial enterprises and format attempts so far are all failures commercially. Meridian loses money and is only kept afloat by his in-laws infusing it with cash. So far the same with MQA. Look at the financial statements.

MQA is a closed proprietary format that essentially offers nothing of value to the consumer. It’s simply another attempt by the big corporations involved to find an avenue to create monopoly rents from selling/streaming music. And to prevent consumers from having actual access to unadulterated master quality files. MQA Ltd, is trying to enlist the record labels/streamers with this vision and then profit from the sale of MQA enabled HW.

The essential claims of MQA are demonstrably false: it’s not superior in SQ, and doesn’t actually offer smaller file size for equivalent resolutions. Many of their secondary claims about the format and how it works have also been shown to be either outright lies, or obfuscations that aren’t much better.

It’s astounding that on a site that claims to be about the “science” of audio, the MQA snake oil is accepted and promoted.

 

[solderdude]

It’s astounding that on a site that claims to be about the “science” of audio, the MQA snake oil is accepted and promoted.

I don't think MQA is accepted and promoted here at all.

MQA is a closed proprietary format that essentially offers nothing of value to the consumer.

On this I guess the vast majority feels the same.

 

[Thomas savage]

Bit short on love here, try and be nice to each other you grumpy old farts.

 

[Music1969]

Paul only allowed the “first unfold” in his streaming module. Nothing else.

No. Their Bridge 2 (made by ConversDigital) does MQA unfolding (leaky upsampling) up to 192kHz.

 

[PierreV]

Bit sort on love here, try and be nice to each other you grumpy old farts.

Yes, always a good idea. But to be honest, while I did not comment about it so far, I am a bit disturbed by the abrupt dismissal of Archimago's stuff.

I found the "but he is not a high-end audiophile" argument really, really weak. The exact same argument was made on the PS Audio forum about this side audience btw. Poor guys looking for bargains and not able to appreciate the greatness of their DACs...

In my opinion, anyone using the argument "I am a high-end audiophile and you are not" has preemptively lost the debate.

There are about 3-4 sites/forums/blogs offering good advice, data sets, rational analysis of stuff. It is a bit sad to see the antipathy...

 

[AudioSceptic]

If Tidal goes out of business, then MQA will die with it and so will all the dire predictions people are making against MQA. And what will kill Tidal would be competition from Amazon (and to some extent Qobuz), not any scare tactics regarding MQA.

This is quite a prediction! You really think that MQA is completely dependent on Tidal?

 

[Rusty Shackleford]

I've been reading the thread, and your comments are repetitive hogwash bud. It sounds like your trying to win a ego battle in your head. Do you have a view or comment on the MQA topic? What's your position?

If pointing out logical fallacies used in defense of MQA is “hogwash,” I’m happy to peddle it.

If it wasn’t already clear, I think MQA’s critics such as Archimago have clearly demonstrated that it’s an inferior, lossy format sold under false pretenses.

 

[AudioSceptic]

It would be as prevalent as MP3 playback is. We got there with WMA with the only company holding out being Apple. They can do the same but their value proposition is too limited. And they may be charging a lot more than we did for WMA. So practical limit to their business plan may be much lower than MP3.

Charging a lot more for what? Sorry but it's not clear to me.

 

[Eirikur]

I found the "but he is not a high-end audiophile" argument really, really weak. The exact same argument was made on the PS Audio forum about this side audience btw. Poor guys looking for bargains and not able to appreciate the greatness of their DACs...

Amir does have a sharp pen at times... and that's how I read it too initially.
The subsequent explanations pertaining to the business side and potential audience for MQA put it in better perspective for me, but his phrasing lends itself to abuse.

 

[Cosmik]

There seem to be two strands to this argument:
1. MQA is designed to 'appropriate' part of the music business; it seeks to make profit from solving a problem that doesn't exist; it could potentially lead to audiophiles losing access to original high res masters, instead having settle for MQA-ed versions; it could lead to an inability to play any high res recordings through open source DSP; it could, potentially, lead to full DRM being enforced through secure hardware; it could lead to audiophiles needing to pay yet again for music they already own.
2. None of the above is likely, so just chill.
2a. And anyway, if something becomes ubiquitous it must be because consumers demanded it.

People arguing for (2) seem to be implying that this negates everything in (1), but it doesn't. It just means that on balance they don't think it's likely. (2a) is very feeble, I think, and we can see in Microsoft's own record breaking fines for anti-competitive practices that ubiquity doesn't always stem from consumer demand. And 'demand' from some people can still lead to the loss of options for everyone else.

 

[AudioSceptic]

In relation to those who demand better than RBCD - very few consumers.

I think it is an insignificant market population but am happy to be disabused of my opinion.

Dolby actually made a significant(measurable) improvement in Compact Cassette performance

Have you forgotten that it made a big difference with studio tape recording before that, and continued to do so until the advent of digital?

 

[PaulD]

Have you forgotten that it made a big difference with studio tape recording before that, and continued to do so until the advent of digital?

Actually, Dolby SR was really good, considerably better than Dolby A (which was *much* better than Dolby B). I remember comparing a Sony 3324 digital machine (16/48 on 1/4" the) to an MCI 24 track machine with 24 channels of Dolby SR - they sounded highly comparable in terms of noise and basic sound quality. Level matched but not blind, and in a studio of course. That was probably towards the end of the 80s... But digital was always going to win, non-linear editing alone would do it, but now we have 24-bit and higher sample rates...

 

[AudioSceptic]

I don't see as even a solution, closer to a nod towards a solution looking for a problem to solve. The bandwidth saving is too minimal to really matter.

And it's moot given the continuing increase in download speeds. In ~20 years we've gone from 64 k dial-up to > 100 M. It's easy to forget that many of us now have internet connections as fast or faster than ethernet 100. We've had similar increases in storage.

A > CD format that makes sense to me would be 20 bit x 64 kHz. 20 bits is close to the best that DACs can manage, and 64 k gives us a Nyquist of 32 k, allowing easy filtering above the audible range. Anything more is a waste of space.

 

[AudioSceptic]

Can't see FLAC in the table.

De-blur is put forward. Has it been defined and heard in listening tests?

Aren't they all FLAC?

 

[sergeauckland]

Actually, Dolby SR was really good, considerably better than Dolby A (which was *much* better than Dolby B). I remember comparing a Sony 3324 digital machine (16/48 on 1/4" the) to an MCI 24 track machine with 24 channels of Dolby SR - they sounded highly comparable in terms of noise and basic sound quality. Level matched but not blind, and in a studio of course. That was probably towards the end of the 80s... But digital was always going to win, non-linear editing alone would do it, but now we have 24-bit and higher sample rates...

Dolby A was 'better' than Dolby B, but only insofar as it was a lot more complex. Dolby A was a three-band process which required tape machines to be immaculately lined up (better than +-1dB on level accuracy and frequency response). This was clearly impossible for consumer machines that were probably never realigned once sold, so Dolby developed a simpler single-band process that wasn't anything as critical on line-up. As the main issue with cassette machines was noise, a single-band HF only band was deemed sufficient as print-through and LF noise was thought less important.

Dolby C was a further development of Dolby B, initially two Dolby B Processors in series, the second with different time constants. Dolby C affected MF and HF, and gave a further 6dB or so of noise reduction (frequency and level dependent) over Dolby B, but required more stringent alignement, which is why so many at the time complained that Dolby C 'killed the music' was 'unmusical' etc etc. Nothing but poor alignment, as well adjusted tape machine was capable of (relatively) superb results with Dolby C.

However, it did take a lot of effort to keep it that way, which is why most of us who were into tape recording in the 1970s & '80s were very quick to adopt digital recording in the 1990s when it became available domestically. No more alignment issues, no more treble crushing issues, no more bass woodles. I still have my first PCMCIA sound card, a Digigram VX Pocket, which still works more than adequately at 44.1 & 48k, 16 & 24 bit. The main problem now is keeping a laptop going with a PCMCIA slot.

S.

 

[LF78]

Apart from all the concerns about DRM and market domination, this is what I understood from a strictly technical standpoint:

• MQA is a lossy codec, more similar to the old ADPCM (adaptive bit depth reduction) than to the "new" perceptual encoders (MP3, AAC, etc...).
• For 44.1 kHz/16 bit standard PCM sources, the codec is actually useless: it degrades original audio (being lossy), while consuming about 2X space compared to lossless.
• For 96 Khz/24 bit high resolution PCM sources, the codec saves about 40% of space. To be honest, if I have to choose a lossy codec I would use the good old AAC at the same bitrate, and I would use good old FLAC for lossless. Much more indipendently tested and measured.
• The minimum phase reconstruction filter used in MQA DACs introduces unwanted imaging artifacts above Nyquist frequency, compared with the standard linear phase brick wall filters commonly used in most competently designed DACs.

Hence from a dry and objective standpoint, using MQA instead of FLAC or AAC (depending on use cases) is similar to preferring a DAC with a worse SINAD (at the same price). Am I wrong?

 

[Dogen]

Apart from all the concerns about DRM and market domination, this is what I understood from a strictly technical standpoint:

• MQA is a lossy codec, more similar to the old ADPCM (adaptive bit depth reduction) than to the "new" perceptual encoders (MP3, AAC, etc...).
• For 44.1 kHz/16 bit standard PCM sources, the codec is actually useless: it degrades original audio (being lossy), while consuming about 2X space compared to lossless.
• For 96 Khz/24 bit high resolution PCM sources, the codec saves about 40% of space. To be honest, if I have to choose a lossy codec I would use the good old AAC at the same bitrate, and I would use good old FLAC for lossless. Much more indipendently tested and measured.
• The minimum phase reconstruction filter used in MQA DACs introduces unwanted imaging artifacts above Nyquist frequency, compared with the standard linear phase brick wall filters commonly used in most competently designed DACs.

Hence from a dry and objective standpoint, using MQA instead of FLAC or AAC (depending on use cases) is similar to preferring a DAC with a worse SINAD (at the same price). Am I wrong?

And don’t forget, MQA cannot be freely digitally processed. So we lose REW, digital bass management, crossover, anything a lot of us like to do with our digital signals. All that is at the pleasure of MQA, using a method they approve. All the open-source digital processing we enjoy today becomes impossible with an MQA encoded file.

 

[georgeT]

And don’t forget, MQA cannot be freely digitally processed. So we lose REW, digital bass management, crossover, anything a lot of us like to do with our digital signals. All that is at the pleasure of MQA, using a method they approve. All the open-source digital processing we enjoy today becomes impossible with an MQA encoded file.

I had no problems using Jriver's convolution engine as a digital crossover using linear phase filters.

 

[Soniclife]

I had no problems using Jriver's convolution engine as a digital crossover using linear phase filters.

There is also no problem using roon DSP on the unfolded data.

 

[Eirikur]

There is also no problem using roon DSP on the unfolded data.

Call me when either Roon or jRiver is available as freeware!