The find_mp3_duplicates PERL script is designed to do one thing: find duplicate audio files, based on the audio content, from two distinct directory trees. It was made to help merge two somewhat-overlapping music collections, but may be adapted for other useful purposes.
It uses the MP3::Info module to extract the audio portion of the file, and then generates an md5 hash of the audio portion. The output assumes that you want to keep the master intact, and remove files from the slave. The basic program flow is:
That said, it, doesn't rely on file name or tag information, which is how most programs find duplicates.
If you make any cool updates I am definitely interested in publishing updated versions--please send them along.