On my drive to the university this morning I started thinking about some really poor quality RealAudio songs I downloaded from the internet when I was on a 28.8k modem and mp3 files were only just become available (i.e. Napster). They were crap, but they began my more alternative music development because the songs were of groups you’d never hear on the radio or on TV.
This got me thinking about lossy compression and copyright. At progressively worse encoding, what point does a song no longer become subject to copyright. Is it based on whether a human can detect the similarity – and is it a human that knows the original by heart, or an impartial observer? Someone who knows the song will be able to pick up underlying patterns of the song, but to an impartial observer who has listened to each version of the song (with a sufficient temporal gap between each listening) they probably wouldn’t.
And if copyright isn’t based on human observers, then eventually songs will overlap with the same information as they are forcibly compressed into smaller sizes.