MUSINGS: On DRM, MQA, the (supposed) Techno-Libertarian opposition, and Honesty... [Plus a quick look at MQA-CD]

Eugène Delacroix - "La Liberté guidant le peuple" (Liberty Leading the People).

Looking around the audiophile press this past couple of weeks, we see the next installment of Jim Austin's exploration into MQA - "MQA, DRM, and Other Four-Letter Words" in the pages of Stereophile.

I asked myself after reading the article, do I and others who share my views on MQA fit the description of being "in a state" about the threat of DRM? And am I and others who oppose MQA the kind of person he generalizes about when he said this?

Because they are a certain kind of person: advocates of open-source, open-standard software. We meet them at the intersection of audiophilia and information technology. These are the Internet libertarians I mentioned earlier: audiophiles whose sensibilities were nourished in the software industry.

For this post, let's think about comments like those above, and consider some peripheral yet important aspects around freedom in audio data file formats, DRM, MQA, and more...

To a certain extent, admittedly the quote above holds some truth speaking mainly for myself of course. I appreciate open-source and open-standards especially for data files because this "opens" the door to compatibility with more hardware and software. Audio software developers can focus energies on creative features, nicer GUI, better playlist management, maybe applying advanced forms of processing like room correction, or even just investing their time into more thorough bug fixing rather than the mundane tasks of file format management which already have been addressed by open-source solutions.

Having said this, while I think data formats lend themselves to be open, I'm certainly not a strict cyber-libertarian though and appreciate all the excellent commercial software out there (if I were a strict free-software user, I'd be using Linux primarily). No doubt the profit motive is a powerful force that leads to excellence and quality all around us and enterprising individuals who can genuinely bring something worthwhile into this world should be rewarded.

Back in October 2016, I said that MQA is a "partially lossy CODEC". Indeed, that is the simple truth. We can actually extend the description a bit more; MQA is a partially lossy bastardized PCM-based CODEC that's typically losslessly compressed in open-standard file formats. For example, the adulterated PCM data (that is MQA) is almost always seen compressed or streamed with FLAC (APE, WV, ALAC, etc. would all be fine). Ironic isn't it that they have basically introduced a proprietary PCM-looking data stream with DRM elements into an open-source lossless file format?!

Open-source software is at the heart of so much we take for granted these days that it would be hypocritical to suggest that any audiophile who uses modern devices would not have "sensibilities" aligned in some way with the open software movement out of gratitude. [Essentially any player or firmware will borrow open-source decoding components and the cottage industry of many audiophile devices like the Sonore Rendu are based on open-source software with customizations.]

As a consumer (and audiophile who appreciates hi-res options), what are some "ideal" characteristics for a data format? Is it not something well-designed that:

1. Includes desired features (eg. tagging, compression, error detection). 

2. Is "open" for compatibility reasons - so it can be played by all of one's devices today and those in the future. 

3. Encourages competition in the marketplace for consumer dollars - no antitrust or  issues with excessive control / monopolization when a format is open and free for any company to support.

And specifically as an audiophile:

4. Uncompromised lossless quality - able to accommodate the different actual resolutions losslessly (16/24-bits, 44.1/48/88.2/96/176.4/192+ kHz). You simply cannot achieve better than the original "studio master" resolutions.

We have had these features and capabilities for more than a decade using standard PCM and FLAC. The standard PCM/FLAC "format" continues to grow in popularity; even Apple is supporting it these days to some extent although sadly not with iTunes natively yet. Another open file format, WavPack 5 (.wv) can even handle both PCM and DSD as discussed last year, including 32-bit integer/floating point resolutions.

So why should audiophile consumers want MQA as opposed to standard PCM when it's now (finally) widely admitted to be lossy and as I discussed previously, consists of elements that can be seen as technically questionable and ultimately diminish sonic resolution compared to a true "original" resolution? No good reason as far as I can tell!

The bottom line is that Jim Austin is yet again arguing for Industry interests which adds cost to the consumer for no real benefit. IMO, MQA Ltd. and the audiophile press tried to "sugar coat" the story of MQA initially by not admitting to the facts: that it has lossy elements, is not a "one-stop-shop" (there are already 3 MQA bitrates out there - 16/44.1 MQA-CD, 24/44.1, and 24/48), tried to razzle-dazzle folks with unsubstantiated filter designs, insisted on audible benefits of questionable "time domain" improvements, and even claimed that they could achieve "better" sound than an original "studio master". This last claim is of course impossible unless they're claiming that the DSP process results in consistently superior sound than what the artists and engineers came out with from the studio. Now, they want to soften the fact that there are "Authentication" elements that obviously allow control over the playback mechanism as "not DRM". As I have said before, I agree that what they're doing today is simply a weak, proprietary scheme that is not copy protection. But with a control mechanism in place, it does open the door to other kinds of creative implementations which would not benefit consumer freedom (again, as discussed previously).

I have in fact over the years on various forums expressed that I can accept DRM and do not insist that all music must be "open" PCM/DSD even though I don't think DRM is a good idea for consumers. I still buy the occasional SACD which is copy protected though circumventable these days, and I accept that DRM is incorporated into the (UHD) Blu-Rays I buy. Yes, "freedom" is good, but when I think about MQA, there is something else that I think gets the informed consumer into more of "a state" than simply DRM.

I believe that if MQA had been more honest, the public reception would not have been so poor.

Imagine if one day the record companies decide that they've had enough of releasing new music in an open file format and that they want to follow the lead of the video industry. DRM in their opinion needs to be implemented for both downloads and streaming contents. So the Industry spokesperson invites the press for a huge announcement one evening over cocktails and hors d'oeuvres (say, at The Shard :-):

Ladies & gentlemen of the press. As you know, we in the music industry feel that piracy has severely damaged profitability. We, the largest music labels in the world have decided to announce a new initiative. Over the next 24 months, we will be gradually releasing all lossless resolution music in Lossless-Audio-Super-Transport (LAST) format.

LAST is a state-of-the-art CODEC and file format that incorporates the latest in extensible features like deep metadata. It will have one of the highest lossless compression ratios. It can handle DSD and PCM audio at full resolutions. Music producers have signed on to include consistent and detailed metadata for user convenience. Production standards have been agreed upon to limit severe dynamic range compression to elevate audio quality. Code for dynamic transcoding for lower bitrate streaming has also been written and will be made available to commercial parties. Finally, the LAST data format will also incorporate HAD (Hyper-Authentication-DRM) to track unauthorized playback and ensure complete data integrity with a blue indicator on your device or DAC. 

We will work with all software and hardware partners to incorporate the necessary decoding in their products. The licensing cost will be extremely reasonable as we want rapid implementation. While we regret that not all devices can be upgraded for LAST compatibility depending on the hardware capabilities, we will make every effort to accommodate legacy devices. 

I will now ask Dr. Haggert to come and discuss the technical aspects with his graphs and comparison charts. Live long and prosper.

An announcement like that, while likely unwelcomed by many would at least be honest and if they actually showed some effort to limit the "Loudness War", maybe then we could at least hope for an attempt at actually improving sound quality (instead of the nonsensical push with copious pseudo-Hi-Res albums that are a waste of storage). No wrestling with the semantics of what is or is not "lossless", or what constitutes "DRM". No specious claims of "better" sound compared to the "studio master". No need to ask the audiophile magazine evangelists to promote the ridiculous hype with weak, subjective hand-waving (further reducing already marginal credibility). And certainly no need to issue any "revolutionary new technology" self-proclamations! Deal head-on with the public openly, "objectively", and most of all honestly.

I hope that magazines and audio writers can appreciate that ultimately, the negative responses toward MQA are multifactorial. "DRM" is not necessarily the main "four-letter word" objection although it practically does affect consumer options and an advanced audiophile's opportunity to apply DSP to the full resolution audio data. It simply comes down to the fact that here in 2018, MQA brings nothing desirable, the technical explanations by the company are unconvincing, the evangelists are not credible, it costs more money, and the attempts at selling this product appear - dare I say - deceptive.

--------------------------------------------

In the evenings, I have been enjoying some quality listening time with my Oppo UDP-205. I see that stock of the product must have sold out and the price on places like Amazon and eBay are getting jacked up. No doubt the device is very well made and at an MSRP of US$1,300, it's not a "steal" but certainly a very reasonable value compared to devices from niche companies. Remember, in the world of audiophilia, compared to luxury purchases such as this likely with zero utilitarian benefit over something one could get "free", the relative value of the Oppo is essentially infinite! :-)

Here's a picture of my TV as I play an "MQA Studio" file off a USB stick on the Oppo (you can see the red USB stick at bottom left inserted into the UDP-205's front USB connector).

As far as I can tell, the device's MQA implementation does not handle the "Rendering" function like the Audioquest Dragonfly products or with the Mytek Brooklyn DAC. Remember that currently, this device is not able to decode MQA as a USB DAC; just when playing back from storage devices plugged into the USB-A ports or playing an MQA-CD. I also notice that it will not see a ripped 16/44.1 "MQA-CD" FLAC file as having the MQA content.

And here's what MQA-CD playback looks like on the Oppo UDP-205 - it's playing the Steve Reich Pulse/Quartet CD:

As you can see, the "original" resolution according to the MQA decoder is 24/88.2kHz. Since we know a thing or two about MQA already, it's no surprise that 16/44.1 CD-resolution data does not have enough bits to lossy encode the upper octave. The question is, then, do we see the use of "leaky" filters with playback? I recorded ~60 seconds of the analogue output of the track "Pulse" in 24/192.

Yes! Indeed... The tell-tale signs of the use of those weak, slow roll-off "leaky" filters could be seen throughout this piece. A couple examples are shown in that screen captures above. The playback through the MQA decoder happened to conveniently add a 22.05kHz tone throughout much of the music even though I don't see this 22.05kHz in the undecoded file. We can see the imaging artifact beyond Nyquist show up like a "mirror image" of the actual signal on either side of the Nyquist frequency. As you know, there is no real high-resolution content with CD-quality (16/44) MQA. The stuff beyond 22.05kHz is clearly nothing like what the "original" 88.2kHz would have looked like. Remember, the fact that there are only 16-bits means there's no lower 8-bits as we typically see in 24-bit MQA to use for the lossy reconstruction of the "unfolded" upper octave.

Watching the FFT display in realtime as the MQA-CD was being played, it looked like there were portions where the artifacts were not as strong. It looks to me like perhaps the decoder was switching between different filter settings in the music - at times like above with a very slow roll-off filter, other times I see what looks like a steeper filter active without as much ultrasonic leakage. For example, the "Quartet" tracks on this MQA-CD appeared to be using a stronger filter setting than the "Pulse" track.

You might be wondering - does MQA-CD sound that different?

Playing the decoded MQA-CD vs. non-decoded FLAC rip over my system at night on the Oppo UDP-205, the simple answer is no. Remember, we're basically looking at 16-bit audio with different digital filter settings and noise-shaped dithering applied in playback.

Obviously, one cannot be expecting a miraculous transformation in fidelity!



Hope you're all enjoying the music... :-)

ADDENDUM:
Okay, just to better clarify what I heard/saw with the Pulse/Quartet MQA-CD and be more specific, here are some graphs of what I meant by seeing the effects from different filters used in the decoding process.

"Linear Phase Fast" refers to one of the 7 selectable digital filters available in the Oppo UDP-205 settings menu.

Notice that the "Pulse" track uses a filter setting with significantly more ultrasonic leakage than the "Quartet" track. The other obvious thing to notice is that over 30 seconds, the decoded/undecoded tracings almost exactly overlapped except above ~18kHz where the frequencies dissociated for both tracks. This actually confirms my subjective impression that whatever "difference" in sound MQA-CD makes with this album is negligible.

If someone claims to be able to hear a significant difference with MQA using a CD like this (perhaps this guy?), it's worth asking the question... Do I have faith in the testimony of Golden Ears? :-)