On the use of the Beta Divergence for Musical Source Separation

Authors

Derry Fitzgerald, Technological University DublinFollow
Matt Cranitch, Cork Institute of TechnologyFollow
Eugene Coyle, Technological University DublinFollow

Document Type

Conference Paper

Rights

This item is available under a Creative Commons License for non-commercial use only

Disciplines

Electrical and electronic engineering

Publication Details

ISSC 2008, Galway, June 18–19.

Abstract

Non-negative Tensor Factorisation based methods have found use in the context of musical sound source separation. These techniques require the use of a suitable cost function to determine the optimal factorisation, and most work has focused on the use of the generalised Kullback-Liebler divergence, and more recently the Itakura-Saito divergence. These divergences can be regarded as limiting cases of the parameterised Beta divergence. This paper looks at the use of the Beta Divergence in the context of musical source separation with a view to determining an optimal value of Beta for this problem. This is considered for both magnitude and power spectrograms. In an effort to avoid potential local minima in the Beta divergence, the use of a “tempered” Beta Divergence is also explored.

Recommended Citation

Fitzgerald, D., Cranitch, M. & Coyle, E. On the use of the Beta Divergence for Musical Source Separation, Irish Signals and Systems Conference, 2008.

Funder

Enterprise Ireland