Underdetermined Instantaneous Audio Source Separation via Local Gaussian Modeling

  • Emmanuel Vincent
  • Simon Arberet
  • Rémi Gribonval
Part of the Lecture Notes in Computer Science book series (LNCS, volume 5441)


Underdetermined source separation is often carried out by modeling time-frequency source coefficients via a fixed sparse prior. This approach fails when the number of active sources in one time-frequency bin is larger than the number of channels or when active sources lie on both sides of an inactive source. In this article, we partially address these issues by modeling time-frequency source coefficients via Gaussian priors with free variances. We study the resulting maximum likelihood criterion and derive a fast non-iterative optimization algorithm that finds the global minimum. We show that this algorithm outperforms state-of-the-art approaches over stereo instantaneous speech mixtures.


Global Minimum Nonzero Entry Active Source Source Separation Blind Source Separation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Zibulevsky, M., Pearlmutter, B.A., Bofill, P., Kisilev, P.: Blind source separation by sparse decomposition in a signal dictionary. In: Independent Component Analysis: Principles and Practice, pp. 181–208. Cambridge Press (2001)Google Scholar
  2. 2.
    Davies, M.E., Mitianoudis, N.: Simple mixture model for sparse overcomplete ICA. IEE Proceedings on Vision, Image and Signal Processing 151(1), 35–43 (2004)CrossRefGoogle Scholar
  3. 3.
    Vincent, E.: Complex nonconvex l p norm minimization for underdetermined source separation. In: Davies, M.E., James, C.J., Abdallah, S.A., Plumbley, M.D. (eds.) ICA 2007. LNCS, vol. 4666, pp. 430–437. Springer, Heidelberg (2007)CrossRefGoogle Scholar
  4. 4.
    Xiao, M., Xie, S., Fu, Y.: A statistically sparse decomposition principle for underdetermined blind source separation. In: Proc. Int. Symp. on Intelligent Signal Processing and Communication Systems (ISPACS), pp. 165–168 (2005)Google Scholar
  5. 5.
    Belouchrani, A., Amin, M.G., Abed-Meraïm, K.: Blind source separation based on time-frequency signal representations. IEEE Trans. on Signal Processing 46(11), 2888–2897 (1998)CrossRefGoogle Scholar
  6. 6.
    Arberet, S., Gribonval, R., Bimbot, F.: A robust method to count and locate audio sources in a stereophonic linear instantaneous mixture. In: Rosca, J.P., Erdogmus, D., Príncipe, J.C., Haykin, S. (eds.) ICA 2006. LNCS, vol. 3889, pp. 536–543. Springer, Heidelberg (2006)CrossRefGoogle Scholar
  7. 7.
    Pham, D.T., Cardoso, J.F.: Blind separation of instantaneous mixtures of non stationary sources. IEEE Trans. on Signal Processing 49(9), 1837–1848 (2001)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Vincent, E., Gribonval, R., Plumbley, M.D.: Oracle estimators for the benchmarking of source separation algorithms. Signal Processing 87(8), 1933–1950 (2007)CrossRefzbMATHGoogle Scholar
  9. 9.
    Pulkki, V., Karjalainen, M.: Localization of amplitude-panned virtual sources I: stereophonic panning. Journal of the Audio Engineering Society 49(9), 739–752 (2001)Google Scholar
  10. 10.
    Vincent, E., Gribonval, R., Févotte, C.: Performance measurement in blind audio source separation. IEEE Trans. on Audio, Speech and Language Processing 14(4), 1462–1469 (2006)CrossRefGoogle Scholar
  11. 11.
    Yılmaz, O., Rickard, S.T.: Blind separation of speech mixtures via time-frequency masking. IEEE Trans. on Signal Processing 52(7), 1830–1847 (2004)MathSciNetCrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2009

Authors and Affiliations

  • Emmanuel Vincent
    • 1
  • Simon Arberet
    • 1
  • Rémi Gribonval
    • 1
  1. 1.METISS GroupIRISA-INRIARennes CedexFrance

Personalised recommendations