[docwra]

You probably want to hash the audio, not the file as a whole?

You could also use existing fingerprint libs, like Acoustid.

Yeh I looked into it a lot but Acoustid is terribly slow and resource intensive and they charge commercial companies so it was a non starter.

We went with 2 tiny file hashes at the start end end of the file that are independent of the tagging and embedded covers.

Initial results are promising but of course it needs clients to write the logic. Zoomplayer has introduced it I think due to a lot of their users missing good tags and feedback has been pretty nice

I don't want to derail this thread, so for more info see here: http://forum.kodi.tv/showthread.php?tid=306073