This was prepared as a final group project for CS51 at Harvard University, Spring '13. It is a music recognition tool inspired by Shazam, Soundhound, and similar technologies.
The core algorithm is modeled after the one outlined by the Shazam founders here:
http://www.ee.columbia.edu/~dpwe/papers/Wang03-shazam.pdf
The app, as completed in 2013, is highly accurate at identifying short (<30 second) samples cut directly from complete audio files previously fingerprinted in the library. Accuracy drops off on samples captured through a microphone.