How does DRM in MQA work?

[wgscott]

In non-specialist terms, can someone explain how Digital Rights Management (presumably some sort of watermark or something akin to it) would work in the MQA codec?

 

Also, if it replaces audible watermarks currently used in some streamed music, could that be a good thing?

 

Let's have a dispassionate, deliberative discussion.

[Brinkman Ship]

Good idea for a thread...

[jabbr]

Since the specification is secret and requires an NDA, there’s no way to really know what it does.

 

Competant formats are open. That’s the only way they can demonstrate their competence.

[Brinkman Ship]

1 minute ago, jabbr said:

Since the specification is secret and requires an NDA, there’s no way to really know what it does.

 

Competant formats are open. That’s the only way they can demonstrate their competence.

...we can good idea of how it works conceptually.

[trappy]

I have noticed that the MQA versions of some of the most obviously watermarked recordings (i.e older DG) do not exhibit watermarking signatures. Especially obvious when both versions are there. I’ll try to find an example later. I’m thinking the Chicago Abbado Mahler 5...

 

No idea how it works...

[jtwrace]

I'm pretty sure it's like this:
 

• Open wallet
• Bend over
• Open wallet
• Repeat 

[Brinkman Ship]

5 minutes ago, jtwrace said:

I'm pretty sure it's like this:
 

• Open wallet
• Bend over
• Open wallet
• Repeat 

...dry, or lubed?

[jtwrace]

3 minutes ago, Brinkman Ship said:

...dry, or lubed?

MQA, we take it anyway we can get it

[esldude]

From Lee Scoggins in the other thread:

 

The DRM code is not for DRM, it's for authentication so they can control the quality of the end product.  It certifies that the resulting file has had the correct MQA filters applied and has been approved by the label/producer/artist or some subset thereof.

 

I don't know if this answers your questions.  It would appear to have the ability to prevent playback as undecoded MQA with the MQA light on.  Some people have shown the light stays on when it should not in some cases anyway.  

 

It would appear to have the potential for abuse in the more common manner of DRM in the future.  Of course MQA people are saying it has never been done and they aren't going to do that. 

[Fokus]

There are two aspects.

 

The first is the 'hard' DRM, a mechanism observed in the code that allows for some audio bits to be swapped, resulting in recognisable but distorted replay unless the user can show the proper key.

AFAIK this is not used today, but it is in there nevertheless.

 

The second is conceptual DRM. The user can never have access to the fully unfolded audio in the digital domain. That means no inspection, and no compatibility with presently-installed (*) digital domain signal processing.

These are two significant restrictions compared to open formats, hence 'managing my rights'.

 

(* Of course MQA invite us to ditch it all and purchase new, MQA-approved processors.)

 

[bambadoo]

http://archimago.blogspot.com/2018/04/musings-on-drm-mqa-supposed-techno.html

 

Some thoughts there. 

[mansr]

The full content of an MQA file cannot be accessed without a licensed decoder. That alone makes it a form of DRM. That the files can be copied in their encoded form is irrelevant. This is true of every software based DRM scheme. Even a DVD can be replicated without decrypting the protected parts. What matters is how the data can be usefully accessed.

 

In the more draconian DRM systems, the decoder contacts a remote server which verifies that the user holds a valid licence and only then supplies the decryption key. The server often imposes severe restrictions such as a limited number of playbacks. If the server becomes inaccessible, possibly permanently, the protected file is rendered useless. MQA is not quite that bad. As it does not "phone home," there can be no per-play checks. Remember, the same is true for DVDs, and everybody agrees that they have DRM.

 

The MQA defenders now point out that their files allow partial playback even without a decoder. This is, however, not the entire truth. Firstly, such playback does not reach even CD quality. Secondly, while reverse engineering the decoder, I found a mechanism that allows the entire file to be encrypted, making it completely unplayable on non-MQA systems. While it is true that no files using this mode have yet been found in the wild, this does not in any way mean there never will. After all, why would the feature be included if there was no intent to use it? There is every reason to believe that if MQA support were to become sufficiently widespread, at least some music would start being distributed with this encryption enabled. If you bought such a file, it would become totally unplayable the day MQA equipment is no longer available, and this day will come. Nothing lasts forever.

[sphinxsix]

6 hours ago, mansr said:

The MQA defenders now point out that their files allow partial playback even without a decoder. This is, however, not the entire truth. Firstly, such playback does not reach even CD quality.

True. My listening test impressions:

A great thread idea! I was asking myself the same question.

[Samuel T Cogley]

18 hours ago, trappy said:

I have noticed that the MQA versions of some of the most obviously watermarked recordings (i.e older DG) do not exhibit watermarking signatures. Especially obvious when both versions are there. I’ll try to find an example later. I’m thinking the Chicago Abbado Mahler 5...

 

I have noticed this as well.  The only favorable thing I can say about MQA.  The MQA'd versions sound much better than the watermarked versions.  Not because of MQA, but because of the absence of the atrocious watermarking.

[Miska]

7 hours ago, mansr said:

The MQA defenders now point out that their files allow partial playback even without a decoder. This is, however, not the entire truth. Firstly, such playback does not reach even CD quality.

 

And that is in many ways similar to the SACD. One could play the RedBook resolution on standard unprotected player, or rip to a computer. But the play the DSD layer, one would need to have licensed hardware. And DSD data could only be sent over protected link to another protected and licensed piece of hardware. Only analog outputs were allowed at full resolution. Digital output over S/PDIF or similar was allowed to be only RedBook quality. And SACD's protections lasted longer than DVD's...

 

MQA is just worse in this respect by not allowing even RedBook quality without licensed and approved decoder.

 

[mansr]

1 minute ago, Miska said:

And that is in many ways similar to the SACD. One could play the RedBook resolution on standard unprotected player, or rip to a computer. But the play the DSD layer, one would need to have licensed hardware. And DSD data could only be sent over protected link to another protected and licensed piece of hardware. Only analog outputs were allowed at full resolution. Digital output over S/PDIF or similar was allowed to be only RedBook quality.

Slight correction: the PS3 outputs SACD as 88.2/24 over Toslink.

[tmtomh]

44 minutes ago, Miska said:

 

And that is in many ways similar to the SACD. One could play the RedBook resolution on standard unprotected player, or rip to a computer. But the play the DSD layer, one would need to have licensed hardware. And DSD data could only be sent over protected link to another protected and licensed piece of hardware. Only analog outputs were allowed at full resolution. Digital output over S/PDIF or similar was allowed to be only RedBook quality. And SACD's protections lasted longer than DVD's...

 

MQA is just worse in this respect by not allowing even RedBook quality without licensed and approved decoder.

 

 

Absolutely yes - and, neither Sony nor anyone else ever tried to claim SACD was not a DRM'd format. And they certainly didn't send people to publicly impugn the integrity who pointed out that the format was DRM'd either.

[Miska]

41 minutes ago, mansr said:

Slight correction: the PS3 outputs SACD as 88.2/24 over Toslink.

 

Ahh, that's something I didn't know! That is quite sensible.

 

All the other SACD players I've seen only output 44.1/16 when playing SACD... Even the 5.1 channel Sony I still have... Which is actually a curious device because of it's Sony DSD processor. It did speaker distance and level adjustments when playing DSD. I even checked the PCB, and for all other channels except front stereo the DAC chip PCM pins are not even connected, so it certainly did that with DSD playback... (it uses Burr-Brown DAC chips)

 

[Ralf11]

54 minutes ago, Miska said:

...

 

MQA is just worse in this respect by not allowing even RedBook quality without licensed and approved decoder.

 

 

 

MP3QA

[mansr]

43 minutes ago, Miska said:

 Ahh, that's something I didn't know! That is quite sensible.

 

All the other SACD players I've seen only output 44.1/16 when playing SACD... Even the 5.1 channel Sony I still have... Which is actually a curious device because of it's Sony DSD processor. It did speaker distance and level adjustments when playing DSD. I even checked the PCB, and for all other channels except front stereo the DAC chip PCM pins are not even connected, so it certainly did that with DSD playback... (it uses Burr-Brown DAC chips)

Filter bypass mode on BB chips uses the DSD pins for PCM input.

[Miska]

35 minutes ago, mansr said:

Filter bypass mode on BB chips uses the DSD pins for PCM input.

 

I'm not sure if the DSD1751 even has such... But I think I checked this loong time ago.

 

Closest datasheet I am now able to find is for DSD1702:

http://www.ti.com/lit/ds/symlink/dsd1702.pdf

 

I'm pretty sure it has remodulator for DSD. They had also PCM to DSD upsampling quite a while ago in their players. Not so different from the good old CS4328 DAC chip. I still have one DAC I built with that chip when it was new. CS4328 datasheet describes:

Quote

The architecture includes an 8× oversampling fil-
ter followed by a 64× oversampled one-bit
delta-sigma modulator. The output from the one
bit modulator controls the polarity of a reference
voltage which is then passed through an ultra-
linear analog low-pass filter. The result is
line-level outputs with no need for further filter-
ing.

 

[Fokus]

8 hours ago, tmtomh said:

And they certainly didn't send people to publicly impugn the integrity who pointed out that the format was DRM'd either.

 

When I published a critical comment on SACD on my website my provider and I got a cease and desist from Sony legal.

[mansr]

11 hours ago, Miska said:

I'm not sure if the DSD1751 even has such... But I think I checked this loong time ago.

 

Closest datasheet I am now able to find is for DSD1702:

http://www.ti.com/lit/ds/symlink/dsd1702.pdf

 

I'm pretty sure it has remodulator for DSD.

It's easy enough to check what's being sent to the chips using a scope or logic analyser.

 

11 hours ago, Miska said:

They had also PCM to DSD upsampling quite a while ago in their players. Not so different from the good old CS4328 DAC chip. I still have one DAC I built with that chip when it was new. CS4328 datasheet describes:

Quote

The architecture includes an 8× oversampling fil-
ter followed by a 64× oversampled one-bit
delta-sigma modulator. The output from the one
bit modulator controls the polarity of a reference
voltage which is then passed through an ultra-
linear analog low-pass filter. The result is
line-level outputs with no need for further filter-
ing.

That sounds like a description of just about any sigma-delta DAC chip.

[Miska]

This is getting "slightly" off-topic...

 

2 hours ago, mansr said:

It's easy enough to check what's being sent to the chips using a scope or logic analyser.

 

I believe I checked with spectrum analyzer when the device was new. Now I tried to check again and the player doesn't recognize DSD layers anymore, thinks those are only CD. Too long in storage or something and the pickup head is dead. Tried cleaning up the lens, etc.

 

That leaves me only with one functional SACD player left - a stereo-only Pioneer one. And I don't know if anybody is selling multi-channel non-PCM converting SACD players anymore now that Oppo also quit that stuff. Lovely part of DRM'ed content, similar thing happened with my HD-DVD discs. Luckily I have some of the SACD's ripped, not so lucky with HD-DVD.

 

Sony also has a PCI card for doing similar plus more advanced DSD-to-DSD processing that is part of the Sonoma recording system:

http://www.superaudiocenter.com/Products.htm

They have their own ASIC they use in the SACD players, etc.

 

2 hours ago, mansr said:

That sounds like a description of just about any sigma-delta DAC chip.

 

Yeah, except this is one of the good old true 2.8 MHz 1-bit thingies. In the output, there's no rise in the noise floor, so the analog filter is nicely designed to match the modulator. There's only small noise bump around 2.8 MHz and notch plus peak in the middle where the 2.8 MHz frequency is. Nowadays I would put some extra analog filter there, as the 2.8 MHz frequency spike is at just around -50 dB. And the performance is quite nice overall.

 

[Hugo9000]

11 hours ago, Fokus said:

 

When I published a critical comment on SACD on my website my provider and I got a cease and desist from Sony legal.

Was it related to a potential or alleged violation of the Digital Millenium Copyright Act?