Separation of singing voice from music accompaniment using matrix factorization method

Songs play an important role in entertainment. An audio signal separation system should be able to identify different audio signals such as speech, music and background noise. In a song the singing voice provides useful information. An automatic singing voice separation system is used for attenuating or removing the music accompaniment. The singing voice becomes a main attractive focus of attention. Singing is used to produces relevant sound with music by the human voice. The paper present the developed algorithm Robust Principal Component Analysis (RPCA) for separating singing voice from music background. This method is a matrix factorization for solving low-rank and sparse matrices. Singing Voice has been effectively separated from the mixture of music accompaniment. Evaluation results of the algorithm shows that this method can achieve around 5.2 dB higher GNSDR (Global Normalized Source to Distortion Ratio) on the MIR-1K dataset. Moreover, we examine the separation results for different values of k.