Originally Posted by
BonzoESC
It's not that intensive; Shazam for iPhone and Android can match a song in a noisy bar with ten seconds of recording out of a collection of 20,000 songs in less than half a second on a normal desktop computer[1], and iTunes Match has a "radio quality" sample that Shazam's algorithm could do in a hundredth of a second. Additionally, they can fingerprint over the whole song (with a similar number of samples as Shazam's ten second fingerprint), allowing for a better match with less randomness.
1:
http://www.ee.columbia.edu/~dpwe/pap...g03-shazam.pdf
I guess it depends on your definition. To match my 15,000 songs, listening to 10 seconds of each song would take 42 hour of processing. And iTunes has
slightly more than 20,000 songs (20 million), and I doubt it's linear. And the database isn't local.
I'm sure Apple has thought of this and suspect that a sample/hash/metadata is probably already embedded in the file so it's nothing more than a signature match with the claimed title/track/artist. I doubt it does a pattern match against the entire catalog if the signature doesn't match - probably just uploads the track for your use but doesn't add it to the master database for other users.
Or something completely different.