Positive Feedback ISSUE 60
march/april 2012



iTunes Plus, Compression, and What It Means to You
by Carl Pultz


The mail list of the Association For Recorded Sound Collections (ARSC), a historically-focused space not often roiled by current events, does occasionally erupt with passionate dialog from an item in the news. Last week, lava was flowing after the announcement by Apple of its new "Mastered for iTunes" program, which comprises a set of software tools to be used in the mastering process to prepare files for submission to iTunes. It is also part of the iTunes Plus marketing campaign, a response to complaints about the compromised quality of iTunes downloads.

The idea is that files can be prepared for iTunes as a recording project is finalized, rather than the older process where iTunes AACs were usually ripped from an album's CD release, out of the control of anybody but Apple. Tunes can now be mixed directly from a high sample-rate, high bit-depth format straight to a 44.1 AAC file that is 32bit floating point, which should be an improvement. Dither can be avoided and more dynamic range retained—at least in principal. Apple doesn't say what bit depth finally emerges from the store, as apparently more encoding still occurs behind the digital curtain.

Another aspect of the tool kit is that the prepped files can be listened to and compared to their source. If the processing does something weird, the engineer has the opportunity to go back and tweak the master in an attempt to improve what comes out of the encoder. It puts some of the control back in the hands of the production team. This sounds great, but…

One of the better articles I've seen that discusses all this was published on Ars Technica. The input from mastering engineers was illuminating. This critical auditioning stage cannot be done with real-time adjustments. The mastered song file has to be processed through Apple's widget, the new AAC file auditioned (after additional processing). Then, if it sounds wrong, adjustments have to be considered, a new master file has to be rendered, processed, evaluated, etc., ad nauseum.

The redoubtable Vlado Meller is quoted as describing the process as "like polishing your Bentley in total darkness, then turning on the lights to see where you missed." His colleague Andy VanDette said, "It was not uncommon to revise tracks three, four, even five times until I got something that compared well with the CD."

Think about this. These engineers make a finished product that strives to reflect what the artist and producer labored to create. Then, they to some degree mutilate that sound, in a struggle to minimize further mutilation.

The mastering tools only run on a Mac platform. Many engineers use Windows, as there is excellent audio software written for that platform. But a more significant issue is the added cost. Mastering costs a lot, and this process is a supreme time sink. Not many projects are likely to get more than a first run through Apple's processor. Many, many projects never have a separate mastering step at all—most classical productions are finished at the mix stage.

What do we get from this wonder? The iTunes Plus files have grown from 128kbps constant bit rate to 256kbps variable bit rate. VBR—yippee! Something we've been able to do with MP3s for about a decade. But Apple is thinking bigger and longer. From the Apple white paper, "Mastered for iTunes technology brief":

"As technology advances and bandwidth, storage, battery life, and processor power increase, keeping the highest quality masters available in our systems allows for full advantage of future improvements to your music. Also, though it may not be apparent because there may not always be a physical, tangible master created in LP or CD format, the iTunes catalog forms an important part of the world's historical and cultural record. These masters matter—especially given the move into the cloud on post-PC devices.1"

Sounds like they're trying to convince us that quality matters, right? This means what you send to Apple today is it. That's the best it's going to be—ever... 44.1kHz/bits? Presumably, we have reached the mountaintop, the only mountain that matters.

Read the white paper. It's full of interesting language, and some misinformation. But then this whole subject has for years been rife with condescension and double-speak, as is the pretty good Ars Technica piece itself. In it, you'll find such statements as this subheading:

"How Apple is battling compression"

They are not; they are assuring its permanence by promoting a tool professionals can spin their wheels and knock their heads against, and a way to make consumers and media reporters think the problem is solved. Or at least as solved as it needs to be, as is made pretty clear by this passage:

"As an audio engineer, VanDette is 'hopeful' hardware and storage capabilities will one day make uncompressed, 24-bit audio a practical standard. For instance, digital music service HDtracks already offers a catalogue of 24-bit audio files at various sampling rates up to 192kHz. But such audiophile quality is only beneficial to those with expensive stereo equipment capable of reproducing the subtle nuances captured in these higher-quality files... Want an uphill battle? Try pushing the bulk of consumers to embrace niche audiophile formats and upgrade to capable equipment.2"

The assumption here is utterly flawed. That day has arrived. What the PC/Mac and file players like Fuze, Cowon, iRiver... anything that hasn't been purposely hampered by commercially-motivated limitations... have offered everybody is inexpensive hardware that will support decent resolution—at least 48/24—with, in my experience, pretty good audio quality; certainly beyond the demands of 256kbps. There is also a glaring false comparison embedded in the argument. The choice is not between obsolete 256kbps/?? and super-rez 192kHz/24. There's a whole lot in the middle!

For the first time, we um that can expand its capability have a nearly universal medias beyond what was once dictated by a rigidly defined format, one that can grow with technology and a person's motivation. We have put in the hands of every preteen in suburban America a playback device vastly superior to what their parents had at that age, and most of their grandparents at any age. Taking full advantage of that doesn't require a new format; mp3 and AAC will support almost any practical data-rate.

But the world's biggest recorded music gatekeeper is hiding behind a false, yet automatically accepted, assertion that common hardware is inadequate. Weirdly, they push the current state of the art off to an uncertain future, severely restrict an upgrade path, neglect an opportunity to educate the public, and discard the chance to genuinely reinforce an image of quality at a cutting edge. Instead, the story is still, "people don't care, so why bother?" Aren't they embarrassed, dragging out such a tired argument?

Not in the least, apparently. Looks like they don't want to support the server space and don't want to challenge the "your whole collection on your iPhone" paradigm. They do want to put this whole resolution thing to bed by dividing off the audiophile audience from the mass audience, continuing the ghetto-iZation of both. Then, yes, they can get on with on-demand streaming over America's third-world-quality Internet, 'cuz it will be just as good.

So here we are, at this moment of opportunity, limited not by a primitive laser mechanism or elusive vinyl pressing quality, but by arbitrarily imposed constraints designed to support the interests of Apple over the interests of its customers, the artists, and the future. 

1. The standard recommended to the sound preservationists of the ARSC is 96kHz. Not all agree that this is always necessary, particularly for low-fidelity recordings, but it does err on the side of caution.

2. This is clearly interpolation by the author, not necessarily the opinion of Andy VanDette.

Carl Pultz is proprietor of Alembic Productions and provides recording and post-production services in Western New York.