Phase-based informed source separation for active listening of music

Nicolas Sturmel (1) , Laurent Daudet (2) , Laurent Girin (1)

1 GIPSA-MAGIC - GIPSA - Machines parlantes, Gestes oro-faciaux, Interaction Face-à-face, Communication augmentée
2 IL - Institut Langevin - Ondes et Images (UMR7587)

This paper presents an informed source separation technique of monophonic mixtures. Although the vast majority of the sepa- ration methods are based on the time-frequency energy of each source, we introduce a new approach using solely phase informa- tion to perform the separation. The sources are iteratively recon- structed using an adaptation of the Multiple Input Spectrogram In- version (MISI) algorithm from Gunawan and Sen. The proposed method is then tested against conventional MISI and Wiener filter- ing on monophonic signals and oracle conditions. Results show that at the cost of a larger computation time, our method outper- forms both MISI and Wiener filtering in oracle conditions with much higher objective quality even with phase quantization.

Mots clés: phase coding; Informed Source Separation

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Nicolas Sturmel, Laurent Daudet, Laurent Girin. Phase-based informed source separation for active listening of music. DAFx 2012 - 15th International Conference on Digital Audio Effects, Sep 2012, York, United Kingdom. pp.n/c. ⟨hal-00807001⟩