That writeup from Xiph is excellent. The comparison with adding ultraviolet and infrared to video makes so much sense. But you're dealing with audiophiles who seriously consider getting hi-end power and ethernet cables. I read somewhere that there was a listening test with speakers connected with hanger wire - and audiophiles couldn't tell.
In the end, it's all physics. I could never hear a quality improvement beyond normal 16bit, 320kbps, no matter how demanding the music.