MQA and DRM

[mansr]

Never was a fan of soap opera dramatics as a way to persuade. Can you share a reference that shows MQA licensing suffers from the same pricing issues as SACD licensing did?

I don't know the cost of "certifying" a DAC, but I assume it's substantial.

[smartin]

I don't know the cost of "certifying" a DAC, but I assume it's substantial.

 

I bet it's actually really cheap... It would be in MQAs interest to have as much market penetration as possible. Much larger revenue stream from content than hardware, and the more hardware that exists, the larger the demand for content. If they continue to push the 'hardware is best' approach, which I'm not so sure they will, they need a lot of hardware out there rather quickly, if they are going to convince content delivery companies to consider MQA (and with articles saying Spotify and Pandora are not going to use MQA, or at least are considering other methods of HiRes delivery, it makes the argument all the more pressing).

 

My guess is they will pull back from the hardware approach (even more than they already have). If the rumors are true, and Roon leverages it's 'intimate knowledge' of roon-ready DACs to a full software decode, it won't be long before that becomes the norm (as has been pointed out, maintaining a database of DAC profiles is easy). Of course that renders hardware based DRM impossible...

 

But, I don't know either =).

[labjr]

I don't believe a full software decode is possible if the DAC isn't known?

[smartin]

I don't believe a full software decode is possible if the DAC isn't known?

 

Well, the software is somewhere, either in the source or the DAC (or somewhere in between...). There's no reason it has to be in the DAC on hardware from a technical standpoint. If Roon/Tidal/Audirvana/etc. know what DAC they are delivering to, they could automatically pick the 'right' filter and perform the full decode. That does remove the possibility of hardware based DRM, but that's the only 'downside' I see, vs. a lot of upside. And to me it seems that would make more sense from a business model standpoint, but, just guessing...

[mansr]

I bet it's actually really cheap... It would be in MQAs interest to have as much market penetration as possible. Much larger revenue stream from content than hardware, and the more hardware that exists, the larger the demand for content. If they continue to push the 'hardware is best' approach, which I'm not so sure they will, they need a lot of hardware out there rather quickly, if they are going to convince content delivery companies to consider MQA (and with articles saying Spotify and Pandora are not going to use MQA, or at least are considering other methods of HiRes delivery, it makes the argument all the more pressing).

 

My guess is they will pull back from the hardware approach (even more than they already have). If the rumors are true, and Roon leverages it's 'intimate knowledge' of roon-ready DACs to a full software decode, it won't be long before that becomes the norm (as has been pointed out, maintaining a database of DAC profiles is easy). Of course that renders hardware based DRM impossible...

 

But, I don't know either =).

Maybe you're right, but I have seen the pricing structure for other technologies (sorry, I can't be specific) in this general space, and it's never cheap.

[smartin]

If I were to take my speculation to the next level, I'd bet MQAs future is being decided behind closed doors right now by Spotify, Pandora, etc. Tidal was a nice appetizer, but nothing compared to what the other big players would bring. If Spotify (and Pandora to a smaller extent) comes to an agreement with the labels to stream some other form of HiRes, MQA will be in deep doo-doo.

 

Content delivers have pricing power now, not the labels, and certainly not MQA.

[labjr]

Well, the software is somewhere, either in the source or the DAC (or somewhere in between...). There's no reason it has to be in the DAC on hardware from a technical standpoint. If Roon/Tidal/Audirvana/etc. know what DAC they are delivering to, they could automatically pick the 'right' filter and perform the full decode. That does remove the possibility of hardware based DRM, but that's the only 'downside' I see, vs. a lot of upside. And to me it seems that would make more sense from a business model standpoint, but, just guessing...

 

Well, if you know the DAC it would be the same as decoding it in the DAC, provided the DAC has the correct filter. Then I think people will realize that MQA is nothing but a filter setting which it basically is. But they are trying to market MQA as an entire process or package which is supposed to be authenticated. At some point calling everything between no decoding and full decoding MQA is getting to seem like just marketing nonsense.

[mansr]

Well, if you know the DAC it would be the same as decoding it in the DAC, provided the DAC has the correct filter. Then I think people will realize that MQA is nothing but a filter setting which it basically is. But they are trying to market MQA as an entire process or package which is supposed to be authenticated. At some point calling everything between no decoding and full decoding MQA is getting to seem like just marketing nonsense.

MQA is DRM wrapped up in marketing nonsense.

[crenca]

If I were to take my speculation to the next level, I'd bet MQAs future is being decided behind closed doors right now by Spotify, Pandora, etc. Tidal was a nice appetizer, but nothing compared to what the other big players would bring. If Spotify (and Pandora to a smaller extent) comes to an agreement with the labels to stream some other form of HiRes, MQA will be in deep doo-doo.

 

Content delivers have pricing power now, not the labels, and certainly not MQA.

 

I for one will be (pleasantly) surprised if Spotify, Pandora, Napster - any and all future streaming 16/44 or greater services (outside of "niche" players that serve mostly Jazz/Classical, etc.) are not MQA (or some other similar scheme).

[scintilla]

MQA is DRM wrapped up in marketing nonsense.

 

No it isn't. Didn't we already do this 12 months ago, where we all got pissy?

 

MQA is advertised as an end-to-end "assurance" and "guarantee" that the audio file is presented as the artist/producer intended it to be heard. A DRM would explicitly inhibit ALL playback outside of user-authenticated systems or inhibit copying of watermarked copies.

 

Let me tell you having worked on patent applications for Microsoft in several technical areas, if they wanted to block your access to a file, they would and could. The OS has some serious cryptography going on which requires attorneys with advanced degrees to take disclosures and write. That isn't what MQA is about from all available evidence, including your tests.

 

MQA advertises and delivers a playable 24/44 or 24/48 file whether you have a decoder or not. It is noise-shaped and delivers objectively a 16/44 cd-quality despite the metadata included in the bit-stream through the noise-shaping. Those files cannot be said to have DRM because they will play on any device. Are they compromised? Maybe, maybe not. Subjective. Look at the noise-floor of an SACD. Not 16/44 all the way from 20-20k, is it?

 

Decoded MQA diverges from the standard by allowing a frequency-extension and some possible optimization through a matched filter. Engineers understand most digital systems as non-optimal at baseband in discussion. As I see it, MQA is truly an extension of HDCD in that the objective was -subjective- improvement of sound-quality through matched filters A/D --> D/A. I prosecuted a couple of those HDCD applications after MS acquired the IP, so I know the specification of the patent applications quite well. The system was never a technical advancement, it was purely subjective.

 

But MQA is advertised as "authenticated" by the source. How do you implement that in the real world? You put in a switch that checks the playback hardware, that's how. If it is MQA-authenticated rendering is allowed on the approved hardware and the optimal reconstruction filter is selected; plus you get a pretty light telling you all is good. Otherwise, the file is played back as an un-authenticated file.

 

No DRM. It is nothing more than an implementation of the end-to-end authentication as advertised. And, there is really no other way to do that than what we have seen already.

 

Occam's razor more often than not works to explain motivations for decisions, technical or business. Here, the simplest answer for what you have observed is most probably the reality. If you were charged with implementing a system for end-to-end file replay authentication with matched filters, how would you implement it? Everything else is pure speculation, probably not grounded in rational thought.

 

Like I said before, chill the fuck out everyone. If its not one conspiracy theory, it's another. None make any sense. Take the simplest answer. Accept it.

 

Although I still want full rendering for my M51 and someone -will- eventually provide it because Bob et al. were stupid for their choices, I do not believe for a second this is about DRM, because it clearly is not.

[PeterSt]

I have been staring at the text in here : https://www.highresaudio.com/en/content/terms

and now suddenly my eye falls on this excerpt :

 

HIGHRESAUDIO offers music downloads only of complete albums in the audio format (codec's) FLAC, ALAC, DSD, DXD-FLAC and MQA in 24-bit (without copy protection).

 

This suggests that it could exist "but not with us". Sadly the semantics don't allow to determine whether this only applies to MQA or to everything.

[PeterSt]

I honestly don't know if the license agreement says "in perpetuity." Someone should purchase an MQA album and ask for the license agreement.

 

https://www.highresaudio.com/en/content/terms

 

The confirmation email of my purchase does not tell anything specifically about the MQA album I bought. It refers to the link above though.

[Fokus]

So far I haven't seen an indication of MQA pushing into the download market, and there appears to be (even) less reason for MQA there than in the streaming segment, since size with downloads matters (even) less than it does with streaming (not that it matters a lot with streaming necessarily, not that there aren't better, open means of making things smaller, etc., etc.).

 

There are two incentives for the download industry to accept MQA.

 

1) It allows them to distribute hires files, without giving customers access to the true original files. Protection of the crown jewels.

 

2) The audiophool community has been made to drool over MQA's alleged deblurring magic and as-the-artist-heard-it nonsense. This includes the less-technically minded consumers as well as the specialist press. This will drive the demand for MQA downloads, and God forbid, MQA discs.

[firedog]

I bet it's actually really cheap... It would be in MQAs interest to have as much market penetration as possible. Much larger revenue stream from content than hardware, and the more hardware that exists, the larger the demand for content. If they continue to push the 'hardware is best' approach, which I'm not so sure they will, they need a lot of hardware out there rather quickly, if they are going to convince content delivery companies to consider MQA (and with articles saying Spotify and Pandora are not going to use MQA, or at least are considering other methods of HiRes delivery, it makes the argument all the more pressing).

 

My guess is they will pull back from the hardware approach (even more than they already have). If the rumors are true, and Roon leverages it's 'intimate knowledge' of roon-ready DACs to a full software decode, it won't be long before that becomes the norm (as has been pointed out, maintaining a database of DAC profiles is easy). Of course that renders hardware based DRM impossible...

 

But, I don't know either =).

 

I've already read about a couple of audiophile (in other words, small) producers who said the expense was definitely a consideration in deciding to forgo MQA certification.

[mansr]

Let me tell you having worked on patent applications for Microsoft in several technical areas, if they wanted to block your access to a file, they would and could. The OS has some serious cryptography going on which requires attorneys with advanced degrees to take disclosures and write. That isn't what MQA is about from all available evidence, including your tests.

 

Huh, lawyers writing code?

 

MQA advertises and delivers a playable 24/44 or 24/48 file whether you have a decoder or not. It is noise-shaped and delivers objectively a 16/44 cd-quality despite the metadata included in the bit-stream through the noise-shaping. Those files cannot be said to have DRM because they will play on any device. Are they compromised? Maybe, maybe not. Subjective. Look at the noise-floor of an SACD. Not 16/44 all the way from 20-20k, is it?

 

Decoded MQA diverges from the standard by allowing a frequency-extension and some possible optimization through a matched filter. Engineers understand most digital systems as non-optimal at baseband in discussion. As I see it, MQA is truly an extension of HDCD in that the objective was -subjective- improvement of sound-quality through matched filters A/D --> D/A. I prosecuted a couple of those HDCD applications after MS acquired the IP, so I know the specification of the patent applications quite well. The system was never a technical advancement, it was purely subjective.

 

I am an engineer, which apparently you are not. The main part of HDCD was a peak extension scheme that allowed squeezing a little more dynamic range into the existing bit depth, kind of like 8-bit A-law encoding. Incidentally, MQA also contains a peak extension component.

 

But MQA is advertised as "authenticated" by the source. How do you implement that in the real world? You put in a switch that checks the playback hardware, that's how. If it is MQA-authenticated rendering is allowed on the approved hardware and the optimal reconstruction filter is selected; plus you get a pretty light telling you all is good. Otherwise, the file is played back as an un-authenticated file.

 

No DRM. It is nothing more than an implementation of the end-to-end authentication as advertised. And, there is really no other way to do that than what we have seen already.

 

Occam's razor more often than not works to explain motivations for decisions, technical or business. Here, the simplest answer for what you have observed is most probably the reality. If you were charged with implementing a system for end-to-end file replay authentication with matched filters, how would you implement it? Everything else is pure speculation, probably not grounded in rational thought.

 

I certainly wouldn't do it the way MQA does.

 

MQA isn't the most potent form of DRM, but it nevertheless is DRM in the sense that access to the full content is restricted to licensed equipment.

[Jud]

Huh, lawyers writing code?

 

He meant lawyers writing disclosures due to the crypto.

[labjr]

DO you think people renting access to music should be able to capture the audio stream?

 

Regardless of what I think, someone will find a way to capture the stream anyway. Guaranteed.

[The Computer Audiophile]

Regardless of what I think, someone will find a way to capture the stream anyway. Guaranteed.

That's usually the case. But things like HDCP has made it much harder and too inconvenient to capture video for most people. I believe the challenge for those implementing such things isn't to make it bulletproof, but to make it hard enough.

[labjr]

That's usually the case. But things like HDCP has made it much harder and too inconvenient to capture video for most people. I believe the challenge for those implementing such things isn't to make it bulletproof, but to make it hard enough.

 

Agreed but there are those who like to prove that it can be done, and are especially motivated when large corporations are trying to keep them out. Everyone doesn't need the knowledge. When software tools become available it suddenly becomes morally okay for everyone else. Like with DVD-FAB and DVD-Audio Extractor etc. BTW, HDCP stripper hardware is available on ebay.

[scintilla]

Huh, lawyers writing code?

What Jud said.

 

I am an engineer, which apparently you are not. The main part of HDCD was a peak extension scheme that allowed squeezing a little more dynamic range into the existing bit depth, kind of like 8-bit A-law encoding. Incidentally, MQA also contains a peak extension component.

I'm an EE and I did coursework in DSP, digital/analog filter design, communications systems and designed and built an exemplary end-to-end baseband PCM system with a switchable lossy channel for use in the school communications labs by future students.

 

Peak extension was almost never used for any HDCD and was not actually the principal purpose of the system. The value in the system was the matched filter sets A/D--> D/A according to the inventors themselves, and in any case the most they could get was ~18 bits, which was practically of almost no utility at the time. MQA is essentially claiming the same ground with 'de-blurring'.

 

 

certainly wouldn't do it the way MQA does.

 

MQA isn't the most potent form of DRM, but it nevertheless is DRM in the sense that access to the full content is restricted to licensed equipment.

Well, there's the rub. If MQA is the only format available for a given title instead of an unmolested hirez version matching the master, then you could argue it is a form of DRM. So far, it looks only to be a bandwidth reduction scheme and will be offered, when at all as a file, as one option. If Warner takes all other hires versions off the table, then it could fairly be characterized as a way to control the absolute resolution of their content in-the-wild, but it still isn't a DRM in that case, it is just a resolution throttle. The definition of DRM is narrow. If I can play it back or copy it and still play it back, it isn't a DRM scheme. Period.

[The Computer Audiophile]

Agreed but there are those who like to prove that it can be done, and are especially motivated when large corporations are trying to keep them out. Everyone doesn't need the knowledge. When software tools become available it suddenly becomes morally okay for everyone else. Like with DVD-FAB and DVD-Audio Extractor etc. BTW, HDCP stripper hardware is available on ebay.

 

Good points. It is kind of a strange thing that people, myself included, think things are OK when software is available to do things like rip Blu-ray discs.

 

I can't imagine being on the other side of this. Produce a product, then have determined people try to crack it just because they don't approve of what you did. It must be aggravating.

[mansr]

The definition of DRM is narrow. If I can play it back or copy it and still play it back, it isn't a DRM scheme. Period.

 

That's your definition and yours alone. Isn't it convenient when you can just redefine words to suit your argument?

[The Computer Audiophile]

I'm an EE and I did coursework in DSP, digital/analog filter design, communications systems and designed and built an exemplary end-to-end baseband PCM system with a switchable lossy channel for use in the school communications labs by future students.

 

Peak extension was almost never used for any HDCD and was not actually the principal purpose of the system. The value in the system was the matched filter sets A/D--> D/A according to the inventors themselves, and in any case the most they could get was ~18 bits, which was practically of almost no utility at the time. MQA is essentially claiming the same ground with 'de-blurring'.

 

 

 

Well, there's the rub. If MQA is the only format available for a given title instead of an unmolested hirez version matching the master, then you could argue it is a form of DRM. So far, it looks only to be a bandwidth reduction scheme and will be offered, when at all as a file, as one option. If Warner takes all other hires versions off the table, then it could fairly be characterized as a way to control the absolute resolution of their content in-the-wild, but it still isn't a DRM in that case, it is just a resolution throttle. The definition of DRM is narrow. If I can play it back or copy it and still play it back, it isn't a DRM scheme. Period.

 

I'd never thought I'd say this, but I pretty much agree with you :~)

 

The definition of DRM is a tough one and I'm glad you've offered your opinion. I don't look at MQA as DRM, but others obviously do. I guess it depends on what the meaning of "is" is :~)

[The Computer Audiophile]

That's your definition and yours alone. Isn't it convenient when you can just redefine words to suit your argument?

 

Good morning Mansr - Do you consider the terms and conditions of purchasing a license to play the music on a CD to be DRM?

[mansr]

Good points. It is kind of a strange thing that people, myself included, think things are OK when software is available to do things like rip Blu-ray discs.

 

I can't imagine being on the other side of this. Produce a product, then have determined people try to crack it just because they don't approve of what you did. It must be aggravating.

 

There's an old guideline for network management (firewalls etc.) that as an administrator you should when at all reasonable provide a safe means for users to do what they want. If you do not, they will find an unsafe way, potentially compromising the entire system (for instance by bridging a mobile device onto the local network). The same thinking applies to restricting usage of digital content. Limit things a little, and people will probably be content with it. Lock it down tightly, and they'll break it completely because you gave them no other choice.