Audio source separation using a redundant library of source spectral bases for non-negative tensor factorization.

Mahmoud Fakhry, Piergiorgio Svaizer, Maurizio Omologo

Browse the full ICASSP paper archive.