Open Access Open Access  Restricted Access Subscription or Fee Access

Speech Enhancement Proposed by a Microphone Array using EM & HMM Model

S.V. Gurao, M.R. Bachute

Abstract


Speech enhancement in a real life working environment is a challenging open problem, which remains unsolved after a long research. The main difficulty towards enhancement is different environmental conditions as well as speech enhancement is processed in real time basis. In this paper a new approach towards speech enhancement is proposed using EM algorithm and Hidden markov model (HMM) model. This unique approach gives information about statistical structure of speech signal using speech model. By using this new technique a speech model is parameterized by the coefficient of the reverberation filter and spectra of the sensor noise signal. In new approach EM algorithm and HMM model work together, so that noiseless speech signal is clearly audible user. The main function of EM algorithm is that to estimate Noise parameter from Hidden variable model and develop a Bayes optimal model of the original speech signal, At the same time reduces the magnitude of noise signal .This new technique of speech enhancement is having better solution over the traditional methods like Spectral subtraction, Array processing ,Noise cancellation. EM algorithm and HMM model work together so we are getting good values of performance parameter as compare to earlier methods. The important parameter for speech enhancement is signal to noise ratio (SNR).So higher the SNR value, noise is less. And result is enhanced speech signal which is noise free.

Keywords


Speech Enhancement, EM Algorithm, HMM Model, Probabilistic Model Approach.

Full Text:

PDF

References


H.Attias, L.Deng, A.Acero, J.C.Platt (2001). Anew method for speech denoising using probabilistic models for clean speech and for noise.Proc.Eurospeech2001W.-K. Chen, Linear Networks and Systems (Book style). Belmont, CA: Wadsworth, 1993, pp. 123–135.

H. Attias, C.E. Schreiner (1998). Blind source separation and deconvolution: the dynamic component analysis algorithm. Neural Computation 10, 1373-1424.

Ephraim, Y. (1992). Statistical model based speech enhancement systems. Proc. IEEE 80(10), 1526-1555.

B.J. Frey, L. Deng, A. Acero, T. Kristjansson (2001). An iterative variational method for removing multiple types of acoustic distortion for robust speech recognition. Proc. Eurospeech 2001.C. J. Kaufman, Rocky Mountain Research Lab., Boulder, CO, private communication, May 1995.

S. Griebel, M. Brandstein (2001). Microphone array speech dereverberation using coarse channel modeling. Proc. ICASSP 2001.

Rabiner, L.R. A tutorial on hidden Markov models and selected applications in speech recognition.Proc.IEEE 77,257–286 (1989)

Dempster, A.P., Laird, N.M., Rubin, D.B. (1977). Maximum Likelihood from Incomplete Data via the EM Algorithm. Journal of the Royal Statistical Society. Series B (Methodological), Vol. 39, No. 1, , pp. 1-38

Redner, R.A., Walker, H.F. (Apr., 1984). Mixture Densities, Maximum Likelihood and the EM Algorithm. SIAM Review, Vol. 26, No. 2., pp. 195-239

M. Askar and H. Derin, A recursive algorithm for the Bayes solution of the smoothing problem, "IEEE Trans. Automatic Control, vol. 26, pp. 558-561, 1981

J. Benesty, J. Chen, and Y. Huang, “Speech Enhancement ” Springer-Verlag, Berlin, Germany, 2005

Philipos C. Loizou ,”Speech Enhancement: Theory and Practice ”,CRC Press. June 7, 2007.

Y. Ephraim and D. Malah, Signal to noise ratio estimation for enhancing speech using the Viterbi algorithm," Technion, EE Pub. No. 489, Mar. 1984

Bilmes, Jeff A. (1998). A Gentle Tutorial of the EM Algorithm and its Application to Parameter Estimation for Gaussian Mixture and Hidden Markov Models. Berkeley, CA: International Computer Science Institute. pp. 7–13.

R. Redner and H. Walker. Mixture densities, maximum likelihood and the em algorithm.SIAM Review, 26(2), 1984.

Breeding, Andy (2004). The Music Internet Untangled: Using Online Services to Expand Your Musical Horizons. Giant Path. p. 128. ISBN 9781932340020.

M. Jordan and L. Xu. Convergence results for the em approach to mixtures of experts architectures. Neural Networks 8:1409–1431, 1996.

M.I. Jordan, Z. Ghahramani, T .S. Jaakk ola,L.K. Saul(1999).An introduction to variational methods in graphical models. Mac hine Learning 37,183-233.

Cohen, On speech enhancement under signal presence uncertainty, Proceedings of the 26th IEEE International Conference on Acoustics Speech, and Signal Processing, ICASSP-01, Salt Lake City, Utah, 7–11 May 2001.

D. Malah, R.V. Cox, A.J. Accardi, Tracking speech-presence uncertainty to improve speech enhancement in non-stationarynoise environments, Proceedings of the 24th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP-99, Phoenix, Arizona, 15–19 March 1999, pp. 789–792.

Marin, J.M., Mengersen, K. and Robert, C.P. "Bayesian modelling and inference on mixtures of distributions". Handbook of Statistics 25, D. Dey and C.R. Rao (eds). Elsevier-Sciences.

M. S. Arulampalam, S. Maskell, N. Gordon and T. Clapp,A tutorial on particle Ølters for online nonlinear/non-Gaussian Bayesian tracking Signal Processing," IEEE Trans. Signal Proc., vol. 50, pp. 174 -188, Feb. 2002


Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.