MQA: A Review of controversies, concerns, and cautions

[Soniclife]

The blue light proves exactly ... nothing.

There was a lovely quote on another forum to the effect of "we now know what the blue light means, it means you have been ripped off".

 

[RayDunzl]

I get the lack of trust of the music biz

Frank discusses the biz way back in the 80's sometime...

 

[Don Hills]

Capabilities? You mean projection of stuff that could happen that don't exist? I read plenty of that from antagonists of MQA. For now, what is sorely missing is them understanding what DRM is. He needs to read the article just posted.

And oh, if he has reverse engineered it, what is the concern about its proprietary nature again?

The code is present in the decoders. All it takes to enable it is to set bit(s) in the MQA control stream. It makes undecoded MQA files effectively unlistenable. The concern about its proprietary nature is that you can't legally decode it without a licensed decoder.

As for understanding what DRM is, the clue is right in the name. It doesn't just mean "unable to make a copy".

 

[Jinjuku]

I also defend Bob Stuart as probably a handful of engineers involved in high-end audio where he actually knows what he is talking about. As such I don't appreciate the rocks people throw at him in order to discredit MQA.

If Mr. Stuart knows what he's talking about then how has the instrumented push back become so effective? MQA has been shown for the house of mirrors that it is. FLAC provides plenty of benefit while imposing none of the negatives.

I also have to disagree with you on the DRM portion. There is a difference between copy protection and DRM. HDCP's ICT is a prime example. For years ICT wasn't switched on. Set manufacturers along with content creators/distributors didn't want ICT enabled. They wanted the masses to get used too, by virtue of consumption, to the higher resolution, 16:9 content.

When the 3rd generation of HD Panels saturated the market guess what happened? All the sudden ICT switch gets flipped and if HD content was going out on an unauthenticated (sound familiar?) output it was down rezzed.

MQA's approach is so ludicrously textbook it's right out of the HDCP consortium playbook. Walmart used to have a music service. It authenticated everything until it didn't and then you couldn't get the content you paid for.

 

[tomelex]

But why would we need one? The real world hasn't changed, just our interface to it. If we add a device to our stereo which compensates the loss of our hearing the stereo will no longer sound anything like real world to us any more, since we listen both to the real world and our stereo with the ears we now have.

I disagree for this reason, since we listen to recordings, in other words, if the recording has the guitar at -5db, but my hearing is down -10 db in that range, then I am not hearing the recording as it was intended. I would rather hear like I did when I was a young one, all the richness, intensity and tones.

 

[amirm]

If Mr. Stuart knows what he's talking about then how has the instrumented push back become so effective?

Let me turn around the question: why on earth is the community going after him instead of the countless people selling junk to audiophiles? Why don't they all get together and go after the people behind those products? After all, few if any of them have the credentials and contributions Bob has.

People really have the wrong guy here. Bob knows more about signal processing than all of these people going after him:

If these people have something to say that is proper and devoid of emotion, then they should write a paper and submit it to AES.

I will say this direct: this dog don't hunt. They should not go after Bob as an individual. He has more than enough qualifications here.

And his contributions in the case of MQA is significant. It is not easy to build a perceptual codec that is backwards compatible with PCM.

Indeed if people want to beat up MQA, they should build their own version of it. If that solution is open and free, then the market can rally around that and MQA will die assuredly.

 

[amirm]

I can't inspect an MQA file in a DAW. My rights are being managed.

If your DAW can't open WMA Lossless because it doesn't have a decode for it, you would say your rights are being managed with respect to that format?

I assume not. Your DAW vendor can contact MQA and get a license to decode those files. That is the reason you can't open the file, not because the file itself has some kind of rights management.

I can't push an MQA file through my DSP amp or my DSP speakers. My rights are being managed.

In the stereophile article they say they allow that now.

You want examples of where your rights are being managed, look to Blu-ray. You can't do anything to the format but to play it. Can't open it with your DAW. Can't make copies. Can't output in uncompressed domain without output protection.

The same is true of any video streaming format like Netflix, Amazon video, etc.

These are proper examples of rights management/copy protection. There is no such mechanism in MQA.

What there is in MQA is authentication of the hardware decoder to make sure it is not bootleg. Every SD card you buy today has the same authentication. They added it because people in China were stealing the firmware and making unlicensed SD cards. The patent holders wanted to get paid so they upgraded the spec with that authentication. That is all that MQA is doing. Should we refuse to buy SD cards because it has the same authentication?

 

[amirm]

The code is present in the decoders. All it takes to enable it is to set bit(s) in the MQA control stream. It makes undecoded MQA files effectively unlistenable.

That is not possible. The baseband signal in undecoded MQA files is in the clear and encapsulated in flac compression. The flac decoder would not pay attention to any such MQA flags and decompresses the music and presents it at 44.1/48 kHz music.

That is the #1 value proposition of MQA: that its baseband signal is in the clear and fully backward compatible with any PCM/Flac playback system. Take that away and there is no reason to go with MQA. All the fanciness is there to encode extra goodies while keeping the baseline as audible PCM with high fidelity no less.

BTW, HDCD did something similar. It required authorized decoders to decode it. Audiophiles embraced it positively. Why wasn't that bad but MQA is?

 

[tomelex]

Amir,

Need a little help here. As audiophiles, with finally the chance to get the master file, at highest current bit depth and rate, that's all we could ask for, it allows for gentler filters etc so technically it should allow also the best reproduction in our homes, despite a lot of it is useless information as far as hearing goes. Memory is cheap and plentiful. Digital transfer rates are fast to our homes.

So, can I ask you simply, what exactly is the problem that MQA is solving for me, as an audiophile, getting that highest resolution master? I don't think I care about backward compatibility, we have been through vinyl, reel to reel, 8 track, cassette, cd and high resolution downloads, and if we can get that master data now, then we don't care about backward compatibility, its a non issue, just add to the list above, after high resolution download, master resolution download...the gold standard.

Kind of like slower speed tape decks, long play records, things that take away from a more perfect replication of the master, I just don't see how MQA improves things for me, the consumer looking to get that gold standard master into my grubby little hands.

 

[Jinjuku]

Let me turn around the question: why on earth is the community going after him instead of the countless people selling junk to audiophiles? Why don't they all get together and go after the people behind those products? After all, few if any of them have the credentials and contributions Bob has.

People really have the wrong guy here. Bob knows more about signal processing than all of these people going after him:

View attachment 12350

If these people have something to say that is proper and devoid of emotion, then they should write a paper and submit it to AES.

I will say this direct: this dog don't hunt. They should not go after Bob as an individual. He has more than enough qualifications here.

And his contributions in the case of MQA is significant. It is not easy to build a perceptual codec that is backwards compatible with PCM.

Indeed if people want to beat up MQA, they should build their own version of it. If that solution is open and free, then the market can rally around that and MQA will die assuredly.

Ok. Then I guess I need a primer on where FLAC streams fall flat for meeting modern criteria of increasing transfer efficiency.

Never mind that you didn't bother to acknowledge that MQA is a form of DRM. Just like ICT from the HDCP protocol is.

It is what it is Amir. None of this has anything to do with Mr. Stuarts intelligence.

 

[amirm]

Amir,

Need a little help here. As audiophiles, with finally the chance to get the master file, at highest current bit depth and rate, that's all we could ask for, it allows for gentler filters etc so technically it should allow also the best reproduction in our homes, despite a lot of it is useless information as far as hearing goes. Memory is cheap and plentiful. Digital transfer rates are fast to our homes.

So, can I ask you simply, what exactly is the problem that MQA is solving for me, as an audiophile, getting that highest resolution master?

It enables you through Tidal and other services like that, to bring > 16 bit audio to streaming market.

The bandwidth hit for high-resolution is huge relative to whatever audible benefit there may be. PCM audio is incredibly wasteful. You double the sample rate for example and that doubles the rate even though what may come with that is just noise.

A Perceptual lossy codec can analyze the high-frequency content and only encode enough bits to represent it, resulting in significant savings in bit rate for the service provider.

 

[amirm]

Never mind that you didn't bother to acknowledge that MQA is a form of DRM. Just like ICT from the HDCP protocol is.

That is not a correct analogy. The only cryptography in MQA is for validating licensed decoders. It is no different than Windows license key.

HDCP on the other hand, fully encrypts the content. Nothing can be seen inside the bitstream unless HDCP allows it. The license key in HDCP not only authenticates the hardware but also allows decryption of content. ICT is a feature that is enforced because the content is encrypted otherwise.

 

[amirm]

Mans has insight in what happens between the various functional blocks of an MQA decoder, but as at least one of these blocks uses encryption he cannot duplicate its function. I.o.w. yes, we know what it does, but no,
we cannot build a free-for-all decoder.

That is because he doesn't understand how cryptography works. When and if MQA becomes popular, its authentication scheme will be torn apart like a paper airplane.

Think about it: Tidal has an MQA decoder. Tidal app is just a Windows app. It can be traced down to each instruction and the authentical keys extracted.

Same can happen to any MQA hardware decoder. None of them have a secure bootloader/flash update. This means their firmware can be extracted, examined, modified, etc.

It really is child's play to break all of this in the hands of the right people.

Now, there are countermeasures for such things but it takes huge amount of expertise which the likes of Microsoft, etc. have. But not MQA. The countermeasures for example require renewable key management. No such feature seems to exist in MQA. Likewise, the code in a Windows app can be obfuscated to make tracing difficult. Again, this is hard to do and requires special IP.

I speak from experience having to defend copy protection in Blu-ray and Microsoft DRM against such breaches.

MQA went and licensed some technology for this. That is NOT a path to success. The initial solution is nothing compared to the ongoing maintenance required to keep a secure system, secure.

 

[March Audio]

Hasn't it been successfully argued that mqa is not transparent to the normal 16bit signal?

Also I have become somewhat unconvinced of the value >16 bit, especially in a streaming environment. Even if you ignore any (unproven) audio benefit, where is the market? A relative tiny number of audiophiles?

Whatever Bob Stuart's previous reputation he has failed to demonstrate mqa efficacy. He has deliberately avoided testing that would do so, and all the while promoting a licencing regime that monotises a solution for a problem that doesn't really exist, both in terms of audio qualty and bandwidth reduction. The cost of bandwidth will continue to fall.

It's all a very cynical exercise from my pov

 

[Werner]

It is not easy to build a perceptual codec that is backwards compatible with PCM.

I disagree. There is little in MQA that took a large effort to develop. From bright bulb idea to working
proof of concept cannot have taken more than a few weeks.

 

[Werner]

Indeed if people want to beat up MQA, they should build their own version of it. If that solution is open and free, then the market can rally around that and MQA will die assuredly.

It exists. FLAC is free and open. And if filter trickery like MQA has any audible benefit then any mastering engineer or music publisher could do this for free at the recording side, with recommendations for specific playback filters at the consumer side. Free. Open.

As for the bandwidth reduction for streaming applications: streaming apps are end-to-end owned by the provider, so they could use any compression or encapsulation protocol they desire. But there is no need whatsoever to extend this to the DAC hardware.

 

[Werner]

That is not possible. The baseband signal in undecoded MQA files is in the clear and encapsulated in flac compression.

The MQA decoder carries code for progressive descrambling of the baseband. Now why would that function be there, if it is never going to be used?

BTW, HDCD did something similar. Audiophiles embraced it positively. Why wasn't that bad but MQA is?

What audiophile embrace is immaterial. HDCD went nowhere. MQA is seeking world dominion. If you don't believe that then
please go and read the public reports at Company House.

MQA must be killed before it does any more damage.

 

[Werner]

(on DSP)
In the stereophile article they say they allow that now.

I said my DSP amp. My DSP speakers.

You believe Stereophile? They copy MQA's doublespeak. Of course DSP and MQA can be combined.

In an MQA-controlled system.

Not in any presently existing DSP system. These would have to be updated, i.e. the manufacturer takes an MQA license
and releases in the field updates, or scrapped.

 

[amirm]

I disagree. There is little in MQA that took a large effort to develop.

Great. When can we have your alternative if it is so easy to do? We can give it away and overnight MQA will be history.

 

[Werner]

Great. When can we have your alternative if it is so easy to do? We can give it away and overnight MQA will be history.

Have you studied MQA in any detail? Do you understand it in depth? If you did you would know how little there is in it.
You don't have to believe me. You (yes you of all people) can do this: pick up the phone and ask JJ.

MQA does not solve any problem. It only brings complexity and cost to the table. There is no problem to solve. Why then would anyone want to spend time building an alternative solution?