I have often sat in coffee shops wondering what method of data extrapolation Shazam used to parse audio to be able to search it's music db. I would think about how I would do it. I use Shazam all the time so it's nice to finally know the basic idea.