Professionally-produced music separation guided by covers

Timothée Gerber (1) , Martin Dutasta (2) , Laurent Girin (2) , Cédric Févotte (3)

1 GIPSA-SAIGA - GIPSA - Signal et Automatique pour la surveillance, le diagnostic et la biomécanique
2 GIPSA-MAGIC - GIPSA - Machines parlantes, Gestes oro-faciaux, Interaction Face-à-face, Communication augmentée
3 LTCI - Laboratoire Traitement et Communication de l'Information

This paper addresses the problem of demixing profession- ally produced music, i.e., recovering the musical source signals that compose a (2-channel stereo) commercial mix signal. Inspired by previous studies using MIDI synthe- sized or hummed signals as external references, we pro- pose to use the multitrack signals of a cover interpretation to guide the separation process with a relevant initializa- tion. This process is carried out within the framework of the multichannel convolutive NMF model and associated EM/MU estimation algorithms. Although subject to the limitations of the convolutive assumption, our experiments confirm the potential of using multitrack cover signals for source separation of commercial music.

Mots clés: music processing; covers; Source separation

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

Timothée Gerber, Martin Dutasta, Laurent Girin, Cédric Févotte. Professionally-produced music separation guided by covers. ISMIR 2012 - International Society for Music Information Retrieval Conference, Oct 2012, Porto, Portugal. pp.n/c. ⟨hal-00807027⟩