Musical source separation is the task of separating out or extracting information regarding one or more musical sources from a mixture. The source separation problem is often cast as a matrix factorization problem where the observed mixture is factored into a product of two matrices. Matrix factorization algorithms such as Independent Component Analysis(ICA) and Non-negative Matrix Factorization (NMF) are two algorithms which have received much attention in the literature. This tutorial style talk will review some of the principles behind these algorithms.
Aanchan Mohan: Tutorial on Matrix Factorization Algorithms for Source Separation
Aanchan Mohan is a PhD student in the Sound Processing and Control Laboratory at McGill University.
November 17, 2009
12:30 - 13:30