To Separate Speech

A System for Recognizing Simultaneous Speech
  • John McDonough
  • Kenichi Kumatani
  • Tobias Gehrig
  • Emilian Stoimenov
  • Uwe Mayer
  • Stefan Schacht
  • Matthias Wölfel
  • Dietrich Klakow
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4892)

Abstract

The PASCAL Speech Separation Challenge (SSC) is based on a corpus of sentences from the Wall Street Journal task read by two speakers simultaneously and captured with two circular eight-channel microphone arrays. This work describes our system for the recognition of such simultaneous speech. Our system has four principal components: A person tracker returns the locations of both active speakers, as well as segmentation information for each utterance, which are often of unequal length; two beamformers in generalized sidelobe canceller (GSC) configuration separate the simultaneous speech by setting their active weight vectors according to a minimum mutual information (MMI) criterion; a postfilter and binary mask operating on the outputs of the beamformers further enhance the separated speech; and finally an automatic speech recognition (ASR) engine based on a weighted finite-state transducer (WFST) returns the most likely word hypotheses for the separated streams. In addition to optimizing each of these components, we investigated the effect of the filter bank design used to perform subband analysis and synthesis during beamforming. On the SSC development data, our system achieved a word error rate of 39.6%.

Keywords

Language Model Binary Mask Perfect Reconstruction Separate Speech Word Error Rate 
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.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Gehrig, T., Klee, U., McDonough, J., Ikbal, S., Wölfel, M., Fügen, C.: Tracking and beamforming for multiple simultaneous speakers with probabilistic data association filters. In: Proc. Interspeech, pp. 2594–2597 (2006)Google Scholar
  2. 2.
    Bar-Shalom, Y., Fortmann, T.E.: Tracking and Data Association. Academic Press, San Diego (1988)MATHGoogle Scholar
  3. 3.
    Van Trees, H.L.: Optimum Array Processing. Wiley-Interscience, Chichester (2002)Google Scholar
  4. 4.
    Hyvärinen, A., Oja, E.: Independent component analysis: Algorithms and applications. Neural Networks 13, 411–430 (2000)CrossRefGoogle Scholar
  5. 5.
    McDonough, J., Kumatani, K.: Minimum mutual information beamforming. Technical Report 107, Interactive Systems Lab, Universität Karlsruhe (August 2006)Google Scholar
  6. 6.
    Kumatani, K., Gehrig, T., Mayer, U., Stoimenov, E., McDonough, J., Wölfel, M.: Adaptive beamforming with a minimum mutual information criterion. IEEE Trans. Audio Speech and Lang. Proc. (to appear)Google Scholar
  7. 7.
    Vaidyanathan, P.P.: Multirate Systems and Filter Banks. Prentice-Hall, Englewood Cliffs (1993)MATHGoogle Scholar
  8. 8.
    de Haan, J.M., Grbic, N., Claesson, I., Nordholm, S.E.: Filter bank design for subband adaptive microphone arrays. IEEE Trans. Speech and Audio Proc. 11(1), 14–23 (2003)CrossRefGoogle Scholar
  9. 9.
    Brehm, H., Stammler, W.: Description and generation of spherically invariant speech-model signals. Signal Processing 12, 119–141 (1987)CrossRefGoogle Scholar
  10. 10.
    Mohri, M., Riley, M., Hindle, D., Ljolje, A., Periera, F.: Full expansion of context-dependent networks in large vocabulary speech recognition. In: Proc. ICASSP, Seattle, vol. II, pp. 665–668 (1998)Google Scholar
  11. 11.
    Mohri, M., Pereira, F., Riley, M.: Weighted finite-state transducers in speech recognition. Computer Speech and Language 16, 69–88 (2002)CrossRefGoogle Scholar
  12. 12.
    Mohri, M., Riley, M.: Network optimizations for large vocabulary speech recognition. Speech Communication 25(3) (1998)Google Scholar
  13. 13.
    Stoimenov, E., McDonough, J.: Modeling polyphone context with weighted finite-state transducers. In: Proc. ICASSP (2006)Google Scholar
  14. 14.
    Stoimenov, E., McDonough, J.: Memory efficient modeling of polyphone context with weighted finite-state transducers. In: Proc. Interspeech (2007)Google Scholar
  15. 15.
    Mohri, M.: Finite-state transducers in language and speech processing. Computational Linguistics 23(2) (1997)Google Scholar
  16. 16.
    Mohri, M., Riley, M.: A weight pushing algorithm for large vocabulary speech recognition. In: Proc. ASRU, Aarlborg, Denmark, September 2001, pp. 1603–1606 (2001)Google Scholar
  17. 17.
    Mohri, M.: Minimization algorithms for sequential transducers. Theoretical Computer Science 234(1–2), 177–201 (2000)MATHCrossRefMathSciNetGoogle Scholar
  18. 18.
    Lincoln, M., McCowan, I., Vepa, J., Maganti, H.: The multi-channel wall street journal audio visual corpus (mc-wsj-av): specification and initial experiments. In: Proc. ASRU, pp. 357–362 (November 2005)Google Scholar
  19. 19.
    Wölfel, M., McDonough, J.: Minimum variance distortionless response spectral estimation, review and refinements. IEEE Signal Processing Magazine 22(5), 117–126 (2005)CrossRefGoogle Scholar
  20. 20.
    Fukunaga, K.: Introduction to Statistical Pattern Recognition. Academic Press, New York (1990)MATHGoogle Scholar
  21. 21.
    Gales, M.J.F.: Semi-tied covariance matrices. In: Proc. ICASSP (1998)Google Scholar
  22. 22.
    Fransen, J., Pye, D., Robinson, T., Woodland, P., Young, S.: Wsjcam0 corpus and recording description. Technical Report CUED/F-INFENG/TR.192, Cambridge University Engineering Department (CUED) Speech Group (September 1994)Google Scholar
  23. 23.
    Deller, J., Hansen, J., Proakis, J.: Discrete-Time Processing of Speech Signals. Macmillan Publishing, New York (1993)Google Scholar
  24. 24.
    Anastasakos, T., McDonough, J., Schwarz, R., Makhoul, J.: A compact model for speaker-adaptive training. In: Proc. ICSLP, pp. 1137–1140 (1996)Google Scholar
  25. 25.
    Uebel, L., Woodland, P.: Improvements in linear transform based speaker adaptation. In: Proc. ICASSP (2001)Google Scholar
  26. 26.
    Wölfel, M.: Mel-Frequenzanpassung der Minimum Varianz Distortionless Response Einhüllenden. In: Proc. of ESSV (2003)Google Scholar
  27. 27.
    Gales, M.J.F.: Maximum likelihood linear transformations for HMM-based speech recognition. Computer Speech and Language 12 (1998)Google Scholar
  28. 28.
    Leggetter, C.J., Woodland, P.C.: Maximum likelihood linear regression for speaker adaptation of continuous density hidden markov models. Computer Speech and Language 9, 171–185 (1995)CrossRefGoogle Scholar
  29. 29.
    McDonough, J., Stoimenov, E., Klakow, D.: An algorithm for fast composition of weighted finite-state transducers. In: Proc. ASRU (submitted, 2007)Google Scholar
  30. 30.
    Simmer, K.U., Bitzer, J., Marro, C.: Post-filtering techniques. In: Branstein, M., Ward, D. (eds.) Microphone Arrays, pp. 39–60. Springer, Heidelberg (2001)Google Scholar
  31. 31.
    McCowan, I., Hari-Krishna, M., Gatica-Perez, D., Moore, D., Ba, S.: Speech acquisition in meetings with an audio-visual sensor array. In: Proceedings of the IEEE International Conference on Multimedia and Expo (ICME) (July 2005)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2008

Authors and Affiliations

  • John McDonough
    • 1
    • 3
  • Kenichi Kumatani
    • 2
    • 3
  • Tobias Gehrig
    • 4
  • Emilian Stoimenov
    • 4
  • Uwe Mayer
    • 4
  • Stefan Schacht
    • 1
  • Matthias Wölfel
    • 4
  • Dietrich Klakow
    • 1
  1. 1.Spoken Language SystemsSaarland UniversitySaarbrückenGermany
  2. 2.IDIAP Research InstituteMartignySwitzerland
  3. 3.Institute for Intelligent Sensor-Actuator SystemsUniversity of KarlsruheGermany
  4. 4.Institute for Theoretical Computer ScienceUniversity of KarlsruheGermany

Personalised recommendations