Marco Arment takes a blind test trying to identify music encoded at three different settings: 128k MP3, 384k MP3, and uncompressed WAV:

NPR’s “hint” to then listen again to the now-labeled uncompressed files, since “it takes time for your ears to adjust”, is, of course, bullshit. If they tell you one is better, you’ll subconsciously believe it’s better. This is why blind tests exist.

I stand by my prior assertion that speakers and headphones matter a lot; amps matter a little; and cables, DACs, higher-than-256k bitrates, and higher-than-44/16 sample rates matter so little that almost nobody can blindly detect differences between them. Allocate your budget accordingly.

Not that it’s surprising, but it looks like there’s plenty of snake oil to go around in the audiophile community.