Creating Music by Listening

A system that analyzes an MP3 collection, models its structural, perceptual, and musical content, and resynthesizes new, personalized, and controlable songs. Developed in the Skeleton environment: a Cocoa library for batched music analysis, visualization, machine learning, and audio synthesis.

Thesis commitee members: Tod Machover, Peter Cariani, Francois Pachet, Julius O. Smith, Barry Vercoe