Over the past week a bunch of clients have asked me about an article from Ars Technica, “Does Mastered for iTunes” Matter to Music“. It’s launched a firestorm, but the findings are pretty obvious: When you leave a little headroom and focus on sound quality instead of loudness, lossy-compressed music sounds a lot better. Similarly obvious, since Apple’s standards lead to better sound, not to mention more and different billable “parts”, professional masterers (myself included!) love this initiative.
Of course The All Night Party can master and proof to these new standards but there are tradeoffs. First, 96 kHz 24 bit mixes are great but unless the mixes came from 96/24 tracks, up sampling just for mix down is more about “marketing bits” than meaningful benefits. Our colleagues at Ultrasuede and other local studios have long recorded and mixed at 48 kHz, 24 bits because it sounds better than 44.1 and works natively with video, and is easy on resources (which keeps costs to artists down). This makes all kinds of sense. Beyond that any benefits depend on the music – high dynamic range classical or jazz featuring acoustic instruments with response far beyond 20 kHz are improved, but that’s not really the case for much amplified pop and rock music performed on contemporary gear. Still, a lot of mastering gear sounds better at 96K/24 bits, so there are a few benefits to up-sampling mixes in mastering (depending on processing chain I’m using). So when I’m processing at that resolution, there’s no reason not to capture/save it that way too. It can’t hurt to give the Apple codecs more information, and some evidence it helps the final sound.
The bigger concern involves how you upload to iTunes. Tunecore and many other digital distros don’t support originals greater than 44.1 kHz with 16 bits. Generally that’s something available to labels with accounts directly at Apple, not individual artists (yet). This negates a benefit mentioned in the Ars piece – Apple will not be able to “upgrade” files on hand to higher quality automagically.
Likewise, as Ian Shepard suggests, the results are not exactly related to fidelity (ie CD = iTunes). The benefits come primarily from not slaughtering the tracks with compression, and not clipping them during ingest. Chicago Mastering’s Bob Weston’s hardcore punk example is telling: he kept the peaks down 3dB and lightened compression, and simply generated a better sounding master. Period. Had he bounced that to CD (rescaled to use all 16 bits) the CD would sound better too. Headroom’s a good thing wherever you have it. But fidelity means input=output, the higher the fidelity, the closer the match (not “sounds better”). As Ian Shepard demonstrates to Ars Technica, any perceptual codec changes the sound. Making it sound euphonically “better” is a change, but also a matter of opinion. Some comments rely on the logic that convinces folks to spend $200 on a special AC cables. Andy VanDette kind of makes this point (maybe unintentionally), blowing holes in those theories: If the residue of a null sounds like music, that’s generally not a good thing. The residue or “difference file” is what’s missing, the what’s “different” between the two versions. Now, you don’t want encoding artifacts and space monkeys there either… ideally you’d prefer something more like analog noise or hiss. The lower the level, the better. When you’re hearing music, you’re hearing sonic and quality differences between the source (CD) and the version (Mastered for iTunes file). In other words, it’s proving it doesn’t sound like the CD. Whether it’s better or worse is a different question, but fidelity is off the table.






