listen, please

We were discussing MP3 compression in my (amazing) recording class, and I got to thinking: has anyone ever made custom perceptual coding models? As in, get a person to come in and do hours of listening tests, then craft the encoder to their hearing? Would this give better results?

Secondly, has anyone ever done “I am sitting in an mp3”? I’ve seen the “I am recording a youtube video” cover of I Am Sitting In A Room, but is there a pure audio equivalent?