John took a look at the code and says he can make it a huge amount more efficient... beware, he'll be all over it :)

Which is why we made it opensourced under the GPL. :)

Which part, BTW? The algorithm for generating the fingerprint of the audio is actually probably going to be outdated relatively soon. We've come up with an even better way to uniquely represent a stream of audio data, and we're in testing phases right now with it.

(O|||||O)