Title :
Unsupervised Single-Channel Separation of Nonstationary Signals Using Gammatone Filterbank and Itakura–Saito Nonnegative Matrix Two-Dimensional Factorizations
Author :
Bin Gao ; Woo, Wai L. ; Dlay, S.S.
Author_Institution :
Sch. of Electr. & Electron. Eng., Newcastle Univ., Newcastle upon Tyne, UK
Abstract :
A new unsupervised single-channel source separation method is presented. The proposed method does not require training knowledge and the separation system is based on nonuniform time-frequency (TF) analysis and feature extraction. Unlike conventional researches that concentrate on the use of spectrogram or its variants, we develop our separation algorithms using an alternative TF representation based on the gammatone filterbank. In particular, we show that the monaural mixed audio signal is considerably more separable in this nonuniform TF domain. We also provide the analysis of signal separability to verify this finding. In addition, we derive two new algorithms that extend the recently published Itakura-Saito nonnegative matrix factorization to the case of convolutive model for the nonstationary source signals. These formulations are based on the Quasi-EM framework and the multiplicative gradient descent (MGD) rule, respectively. Experimental tests have been conducted which show that the proposed method is efficient in extracting the sources´ spectral-temporal features that are characterized by large dynamic range of energy, and thus leading to significant improvement in source separation performance.
Keywords :
audio signal processing; channel bank filters; feature extraction; gradient methods; matrix decomposition; signal sources; source separation; time-frequency analysis; Itakura-Saito nonnegative matrix; MGD rule; TF analysis; audio signal; convolutive model; feature extraction; gammatone filterbank; monaural mixed signal; multiplicative gradient descent rule; nonstationary source signal; nonuniform time-frequency; quasi-EM framework; signal separability; single-channel source separation method; spectral-temporal features; spectrogram; two-dimensional factorization; Algorithm design and analysis; Hidden Markov models; Source separation; Spectrogram; Speech; Time frequency analysis; Training; Gammatone filterbank; Itakura–Saito divergence; matrix factorization; nonstationary source separation;
Journal_Title :
Circuits and Systems I: Regular Papers, IEEE Transactions on
DOI :
10.1109/TCSI.2012.2215735