The Journal of the Acoustical Society of Korea

The Acoustical Society of Korea (한국음향학회)
권호 표지

A Frequency-Domain Normalized MBD Algorithm with Unidirectional Filters for Blind Speech Separation

  • Kim Hye-Jin (Department of Electronic Engineering, Paichai University) ;
  • Nam Seung-Hyon (Department of Electronic Engineering, Paichai University)
  • Published : 2005.06.01
Download PDF
⟨ Previous Next ⟩

Abstract

A new multichannel blind deconvolution algorithm is proposed for speech mixtures. It employs unidirectional filters and normalization of gradient terms in the frequency domain. The proposed algorithm is shown to be approximately nonholonomic. Thus it provides improved convergence and separation performances without whitening effect for nonstationary sources such as speech and audio signals. Simulations using real world recordings confirm superior performances over existing algorithms and its usefulness for real applications.

Keywords

References

  1. S. C. Douglas, A. Cichocki, S. I. Amari., and H. H Yang, 'Novel on-line adaptive learning algorithms for blind deconvolution using the natural gradient approach,' Proc. IEEE 11th IFAC Symposium on System Identification, kitakyushu, Japan, 1057-1062, 1997
  2. T. W. Lee, A. Bell, and R. Orglmeister, 'Blind source separation of real world signals,' Proc. IEEE Int. Conf. Neural Networks, Houston, 2129-2135, June 1997
  3. C. Fancourt and L. Parra, 'Coherence function as a criterion for blind source separation,' Proc. IEEE International Workshop on Neural Networks and Signal Processing, 303-312, 2001
  4. S. I. Amari, T. P. Chen, and A. Cichocki, Nonholonomic orthogonal leaming algorithms for blind source separation, (Neural Computation), 12, 2000
  5. X. Sun and S. Douglas, 'A natural gradient convolutive blind source separation algorithm for speech mixtures,' Proc. Int. Workshop on Independent Component Analysis, 2001, San Diego, California, 59-64, 2001
  6. R. H. Lambert, Multichannel blind deconvolution: FIR matrix algebra and separation of multipath mixtures, (Ph.D dissertation, University of Southern California, Los Angels, CA. 1996)
  7. S. C. Douglas and S. Haykin, 'Relationship between blind deconvolution and blind source separation,' Unsupervised Adaptive Filtering: 2 Blind Deconvolution, Ed, S. Haykin, Wiley, 113-145, 2000
  8. S. C. Douglas, C. A. Cichocki, and S. Amari, 'Self-whitening algorithm for adaptive equalization and deconvolution,' IEEE Trans. Signal Processing, 47, 1161-1165, 1999 https://doi.org/10.1109/78.752617
  9. S. Haykin, Adaptive Filter Theory, 4th ed., Prentice-Hall, 2002
  10. L. Zhang, A. Cichocki, and S. I. Amai, 'Geometrical structures of FIR manifold and their application to multichannel blind deconvolution,' Proc. IEEE Workshop on Neural Networks for Signal Processing, Madison, Wisconsin, 303-312, 1999
  11. A. Cichocki and S. I. Amari, Adaptive Blind Signal and Image Processing: learning algorithms and applications, (Wiley, 2002)
  12. E. R. Ferrara, 'Fast implementation of LMS adaptive filter,' IEEE Trans. on Acoustics Speech and Signal Processing, 28, 1980
  13. K. Na, S. Kang, K. Lee, and S. Chae, 'Frequency domain implementation of block adaptive filters for ica-based multichannel blind deconvolution,' Proc. IEEE Conf. on Acoustics, Speech and Signal Processing, 1999
  14. M. Joho and P. Schniter, 'Frequency domain realization of a multichannel blind deconvolution algorithm based on the natural gradient,' Proc. Int Workshop on Independent Component Analysis and Signal Separation, 2003
  15. L. Parra and C. Spence, 'Convolutive blind source separation of nonstationary sources,' IEEE Trans. Speech and Audio Processing, 8 (3), 320-327, May 2000 https://doi.org/10.1109/89.841214
  16. http://inc2,ucsd2.edu/~tewon/