Perceptual Single-Channel Audio Source Separation by Non-negative Matrix Factorization

Kirbiz S., Günsel Kalyoncu B.

IEEE 17th Signal Processing and Communications Applications Conference, Antalya, Türkiye, 9 - 11 Nisan 2009, ss.654-657 identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Antalya
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.654-657
  • İstanbul Teknik Üniversitesi Adresli: Evet

Özet

This paper proposes a single-channel audio source decomposition method that integrates perceptual quality criteria into source separation. Unlike the existing methods, the proposed method applies a perceptually weighted non-negative matrix factorization on log-frequency spectrogram of the mixed signal. The weights are adaptively calculated for each critical band based on a perceptual model described by ITU-R BS. 138 7 perceptual quality standard. It is shown that the proposed adaptive weighting scheme significantly improves the quality of audio sources estimated by minimizing the weighted divergence between the observed log-frequency spectrogram and the model.