Since my previous article on Apple's Mastered for iTunes program, more information has come to light. We now know what "Mastered for iTunes" actually means! NPR sums it up best; check the addendum at the end of this article.
"...I spoke again with Bob Ludwig, the mastering engineer quoted in the story, who has submitted "Mastered for iTunes" tracks to Apple. He says the company is simply providing mastering engineers with tools that allow them to see how songs mastered at 24 bits will clip (that is, distort audibly) when they go through the standardized AAC encoding process. The uncompressed files are then submitted to iTunes, which creates lossless versions before encoding the songs as 256 kpbs AAC files for sale in the iTunes store.
...Why is this significant? Because the fact that Apple retains the lossless versions of the high-quality studio masters means that iTunes, at any time it decides to, can begin selling higher-quality encodes, or even lossless files."
Ars Technica chimed in with their opinion. With the aid of some professional audio engineers, they concluded that "Mastered for iTunes" can make a positive difference, though it should be noted that not all of the audio engineers agreed with each other.
Is Everybody Missing The Point? Kind of. A lot of the discussion has centered around the fact that it's almost physically impossible for us to hear the difference between 24-bit and 16-bit audio, or 96khz and 44.1khz audio. While true, that misses the real point of high resolution audio.
Whenever audio is transformed, data can be lost. It's just a mathematical reality. By default, audio goes through quite a few steps in the pipeline before making it to your ears. iTunes' volume control, its Sound Check and Sound Enhancer features, and the built-in equalizer all play a role. So do the volume controls built into Windows/OSX, as well as other sound "enhancements" performed by your audio device.
With high-resolution audio, there's simply more room for error - all those little rounding errors likely won't add up to something your ears can detect. However, with 44.1khz/16bit ("CD-quality") audio, there's not much room for error: 44.1/16 is just good enough to cover the range of normal human hearing, and excessive audio processing quickly adds up to something our ears can detect.
In many ways, it's exactly like working with lossy JPEGs. JPEGs are fine for viewing and can be nearly indistinguishable from uncompressed master photos, but once you start editing JPEGs extensively all of the artifacts pile up pretty quickly.