In an article published in the March 2018 Stereophile, I wrote that critics have been attacking MQA, the audio codec developed by J. Robert Stuart and Peter Craven, by accusing it of being lossy. The critics are right: MQA is, in fact, a lossy codec—that is, not all of the data in the original recording are recovered when played back via MQA—though in a clever and innocuous way. For MQA’s critics, though, that’s not the point: They use lossy mainly for its negative emotional associations: When audiophiles hear lossy, they think MP3.
Lately, another word—actually, an initialism—is being used in much the same way: DRM, which stands for Digital Rights Management. Like lossy codecs, DRM has a long, checkered history in and around audio. The “Red Book” set of specifications, which defines the Compact Disc, includes no DRM. At the beginning of the 1980s, when Sony and Philips came up with the CD, the Internet wasn’t yet a gleam in Vint Cerf’s eye, and home CD burners were still a decade or so away. Record companies have made several attempts at adding DRM to CD (footnote 1) and to other forms of digital audio, with consistently bad results: Think audible watermarks, Sony root-kits, and that iTunes DRM that, early on, made downloaded music unplayable on any device not made by Apple.
What is Digital Rights Management, anyway? Good question. One problem with debates about DRM—including the one about MQA—is that there’s no official or widely accepted meaning. For some people, DRM means copy protection, like that on DVDs and SACDs: you can’t legally make a digital clone of the disc you’ve just bought or borrowed. Others think that online license “keys,” like those used for some video games or Microsoft software, are clear examples of DRM. But one vocal group—the Internet libertarians, the open-source–software crowd—takes a more extreme position: Anything that’s not 100% accessible for them to copy, and manipulate—every bit and every line of code—is, to them, subject to DRM by definition.
With such a wide range of definitions of digital rights management—and no technical definition or legal standard—how can we sort out the differences?
Here’s one way: Forget the phrase digital rights management. Instead, consider, very simply, which consumer rights, if any, MQA tries to manage. If we leave the loaded phrase out of the discussion, maybe we can avoid acrimony and keep things calm and rational.
What rights does MQA manage? First and most obviously, you can’t build your own MQA-capable DAC, at least not without paying a licensing fee. Decoding of MQA requires a licensed decoder, which operates using proprietary technology. In practical terms, this is little different from playing a stereo LP—to do that, you need a turntable equipped with a stereo cartridge. You can hear the music with a mono cartridge, but you’ll miss out on certain features of the sound. The real, substantive difference between MQA and the stereo LP is a legal agreement and a licensing fee. Also, currently, you can’t transcode an MQA file into another format (eg, FLAC), at least not while preserving its full resolution. The proprietary status of MQA means that third-party developers can’t freely develop their own MQA-enabled apps, at least not without signing an agreement with MQA Ltd.
If that makes it sound as if MQA’s limitations are more legal than practical, then, well, I agree with you. It’s not open-source. It’s not an open standard. MQA is proprietary. But for a proprietary technology, it’s remarkably free of intrusive controls.
What other consumer rights does MQA manage? You can’t inspect the contents of an MQA file (beyond the undecoded version, which typically is just FLAC), and you can’t alter an MQA file at will—again, beyond that open, undecoded version). One oft-mentioned issue is room correction via digital signal processing (DSP)—right now, you can’t do it with an MQA-enabled DAC. (It’s possible but not simple—see diagram). This is not, however, an attempt to manage consumers’ rights so much as an attempt to balance one of MQA’s key objectives—authentication—against consumer control. How do you allow consumers to alter the music signal while assuring that what’s delivered is consistent with the intent of the original artist(s) and/or mastering engineer(s) and/or record label?
How an MQA-equipped DAC can implement volume adjustment and equalization: A full MQA decoder comes in two parts: The Core Decoder (which authenticates and losslessly recovers the encapsulated signal), and an MQA Renderer (which is matched to the associated D/A stage to deliver the correct analog output). The paths shown in blue should be bit-accurate (lossless) to retain the full end-to-end performance. Digital volume/leveling can be implemented in the Renderer. This diagram shows how to include additional processing, such as EQ. The Core Decoder needs to pass instructions to the Renderer; these instructions can follow a side path or be temporarily buried in the audio datastream fed to the Renderer.
Anyway, this latter limitation is starting to disappear. Certain MQA-enabled digital loudspeakers can already combine MQA and DSP room correction, and apparently there’s more to come. “Room Correction–enabled [DAC] implementations are coming to market,” Bob Stuart told me in an e-mail, “but done in a way that supports provenance and/or artist intention.” You can do it, or soon will be able to, but maybe not exactly the way you want to. We’ll see.
What other limitations does MQA impose? I’ve thought hard, but I can’t think of any—despite the emergence (or reemergence) of two documents viewed by MQA opponents as smoking guns. First is an MQA patent application that documents an invention “whereby a representation of an original PCM signal may be reversibly degraded in a controlled manner.” (footnote 2)
Sounds like DRM—but not only was this DRM technology never implemented in MQA; the application was abandoned, as anyone can see by visiting the Public Patent Application Information Retrieval portal.” (footnote 3) “We abandoned [it] in all territories which can be verified,” Stuart told me in an email, with characteristic precision. “Nothing contained in US20160005411 happens in MQA.”
Next there’s the Utamico “case study” that can be found online, which describes that company’s partnership with MQA in developing technology that includes “digital signatures” and “cryptographic keys.” Such things could be used to control access to MQA files and limit consumer rights—but it’s clear from even a cursory read of the case study that the goal is authentication, not access, use, or copy protection. “To ensure the integrity of the artist’s music from the original source to the end listener, MQA needed a solution for securely signing the music file, to ensure cryptographically that what the listener hears is what the artist approved,” the case study says. In 2016, Stuart told Audiostream’s Michael Lavorgna that in MQA, “Everything that looks like security is authentication.”
What about the future? Would MQA ever add features that further limit consumer choice? I posed that question to Stuart; a long exchange of e-mails ensued, ending with this:
Austin: It sounds as though you are saying that while MQA files or streams could in future be delivered via copy-protected media or transmission modes, that is beyond MQA’s control and has nothing to do with MQA per se. MQA itself has no DRM capabilities and will not add any in future.
Stuart: Exactly.
In a follow-up email, Stuart wrote that MQA “takes a strong stance against DRM. We don’t believe in it for music distribution, we don’t provide for it now or in the future.”
Why, then, are some people in a state about the threat of MQA’s supposed DRM? 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. Chris Hermansen, who writes the “Open Music” column at OpenSource.com, expressed their perspective concisely: “[H]aving had for several years the ability to acquire music in open formats that do not interfere with our ability to select software or equipment to play them back, [with MQA] we are being invited to return to the old days where the music industry controls our playback devices. I don’t think this is a good thing.”
That’s the MQA/DRM debate in a nutshell—although, in the title of his column, Hermansen does implicitly acknowledge that MQA is not DRM: “Why the Proprietary MQA Music Encoding System Is Better than DRM, but Still Not Good.”
I contribute to Wikipedia, and even used Linux for a while, but I also enjoy my iPhone and my MacBook Pro—which is to say that I’m okay with proprietary technology. I also watched the collapse of the record business, which coincided with and appears to have been the direct result of open, readily shared technology. It seems I do not fear the prospect of returning audio audio files to the proprietary control of their creators.
Initiaiisms aside, if you share Hermansen’s values, you may have issues with MQA. If your values are closer to mine, you shouldn’t care much about the “DRM” aspect of the great MQA debate.
Next month: I examine MQA’s technical behavior in the light of recent work on post-Shannon sampling theory.
Footnote 1: Technically, any variant of the Compact Disc that includes DRM isn’t a CD at all; it can’t be labeled or sold as a Compact Disc, and can’t bear the Compact Disc logo.
Footnote 2: See John Robert Stuart, Richard J. Hollinshead, Peter Graham Craven, Malcolm Law, “Versatile Music Distribution,” US Patent 20160005411 A1, 7 January 2016.
Footnote 3: Go to https://portal.uspto.gov/pair/PublicPair and search for Application Number 14/765,916.
Click Here: cheap nsw blues jersey