A local discontinuity based approach for monaural singing voice separation from accompanying music with multi-stage non-negative matrix factorization

This paper presents a new method to improve a multi-stage non-negative matrix factorization (NMF) based algorithm for separating singing voice from accompanying music in single channel recordings. In the proposed method, local spectral and temporal discontinuity measures are used to refine the vocal and music components obtained by the baseline NMF algorithm. The effectiveness of the proposed algorithm is demonstrated using the MIR-1K data set.