Music Decomposition

 

James Owers

 
We propose to design generative models for music synthesis with the goal of repairing damaged audio, generating new instruments, and obtaining a general method for describing timbre.

Supervisors: Amos Storkey & Mark Steedman