Or simpler still - the iTunes application will generate fingerprint of the whole audio in each file and send it along with whatever metadata is available. Then serverside the fingerprint and metadata are analyzed and if a certain criteria for a match are not met, iTunes is prompted to upload the file.