Consistent Wiener Filtering: Generalized Time-Frequency Masking Respecting Spectrogram Consistency
Wiener filtering is one of the most widely used methods in audio source separation. It is often applied on time-frequency representations of signals, such as the short-time Fourier transform (STFT), to exploit their short-term stationarity, but so far the design of the Wiener time-frequency mask did not take into account the necessity for the output spectrograms to be consistent, i.e., to correspond to the STFT of a time-domain signal. In this paper, we generalize the concept of Wiener filtering to time-frequency masks which can involve manipulation of the phase as well by formulating the problem as a consistency-constrained Maximum-Likelihood one. We present two methods to solve the problem, one looking for the optimal time-domain signal, the other promoting consistency through a penalty function directly in the time-frequency domain. We show through experimental evaluation that, both in oracle conditions and combined with spectral subtraction, our method outperforms classical Wiener filtering.
KeywordsPenalty Function Source Separation Consistency Constraint Noise Power Spectrum Spectral Subtraction
Unable to display preview. Download preview PDF.
- 2.Vincent, E., Jafari, M.G., Abdallah, S.A., Plumbley, M.D., Davies, M.E.: Probabilistic modeling paradigms for audio source separation. In: Machine Audition: Principles, Algorithms and Systems. IGI Global (to appear)Google Scholar
- 3.Le Roux, J., Ono, N., Sagayama, S.: Explicit consistency constraints for STFT spectrograms and their application to phase reconstruction. In: Proc. SAPA, pp. 23–28 (September 2008)Google Scholar
- 4.Le Roux, J., Kameoka, H., Vincent, E., Ono, N., Kashino, K., Sagayama, S.: Complex NMF under spectrogram consistency constraints. In: Proc. ASJ Autumn Meeting, (2-4-5) (September 2009)Google Scholar
- 7.Lee, D.D., Seung, H.S.: Algorithms for non-negative matrix factorization. In: Proc. NIPS*2000, pp. 556–562. The MIT Press, Cambridge (2001)Google Scholar
- 8.Vincent, E., Gribonval, R., Plumbley, M.D.: BSS Oracle Toolbox Version 2.1, http://bass-db.gforge.inria.fr/bssoracle/
- 9.Vincent, E., Sawada, H., Bofill, P., Makino, S., Rosca, J.P.: First stereo audio source separation evaluation campaign: Data, algorithms and results. In: Proc. ICA, pp. 552–559 (September 2007)Google Scholar
- 10.Araki, S., Ozerov, A., Gowreesunker, V., Sawada, H., Theis, F., Nolte, G., Lutter, D., Duong, N.Q.: The 2010 signal separation evaluation campaign (SiSEC 2010) –Part II–: Audio source separation challenges. In: Proc. LVA/ICA (2010)Google Scholar