White and Color Noise Cancellation of Speech Signal by Adaptive Filtering and Soft Computing Algorithms

In this study, Gaussian white noise and color noise of speech signal are reduced by using adaptive filter and soft computing algorithms. Since the main target is noise reduction of speech signal in a car, ambient noise recorded in a BMW750i is used as color noise in the applications. Signal Noise Ratios (SNR) are selected as +5, 0 and -5 dB for white and color noise. Normalized Least Mean Square (NLMS), Recursive Least Square (RLS) and Genetic Algorithms (GA), Multilayer Perceptron Artificial Neural Network (MLP ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS) are used as adaptive filter and soft computing algorithms, respectively. 5 female and 5 male speakers have been chosen as Speech data from database of Center for Spoken Language Understanding (CSLU) Speaker Verification version 1.1. Noise cancellation performances of the algorithms have been compared by means of Mean Squared Error (MSE). Also processing durations (second) of the algorithms are determined for evaluating possibility of real time implementation. While, the best result is obtained by GA for noise cancellation performance, RLS is the fastest algorithm for real time implementation.

[1]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[2]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[3]  Rafik Goubran,et al.  Acoustic noise suppression using regressive adaptive filtering , 1990, 40th IEEE Conference on Vehicular Technology.

[4]  H. Kwan,et al.  Speech enhancement using adaptive neuro-fuzzy filtering , 2005, 2005 International Symposium on Intelligent Signal Processing and Communication Systems.

[5]  Sam Kwong,et al.  Genetic algorithms: concepts and applications [in engineering design] , 1996, IEEE Trans. Ind. Electron..

[6]  Mohammad Bagher Menhaj,et al.  Training feedforward networks with the Marquardt algorithm , 1994, IEEE Trans. Neural Networks.

[7]  J. Nazuno.  Haykin, Simon. Neural networks: A comprehensive foundation, Prentice Hall, Inc. Segunda Edición, 1999 , 2000 .

[8]  Manolis Papadrakakis,et al.  Learning improvement of neural networks used in structural optimization , 2004 .

[9]  Amir Hussain,et al.  Binaural sub-band adaptive speech enhancement using artificial neural networks , 1998, Speech Commun..

[10]  Eduardo Lleida,et al.  Speech reinforcement system for car cabin communications , 2005, IEEE Transactions on Speech and Audio Processing.

[11]  Simon Haykin,et al.  Adaptive Filter Theory 4th Edition , 2002 .

[12]  Sam Kwong,et al.  Genetic algorithms: concepts and applications [in engineering design] , 1996, IEEE Trans. Ind. Electron..

[13]  Theodore S. Rappaport,et al.  Evaluation of several adaptive algorithms for canceling acoustic noise in mobile radio environments , 1991, [1991 Proceedings] 41st IEEE Vehicular Technology Conference.

[14]  Jyh-Shing Roger Jang,et al.  ANFIS: adaptive-network-based fuzzy inference system , 1993, IEEE Trans. Syst. Man Cybern..

[15]  R. R. Saldanha,et al.  Improvements in genetic algorithms , 2001 .

[16]  B. Yegnanarayana,et al.  Artificial Neural Networks , 2004 .

[17]  S. Dhanjal.  Artificial neural networks in speech processing: problems and challenges , 2001, 2001 IEEE Pacific Rim Conference on Communications, Computers and Signal Processing (IEEE Cat. No.01CH37233).

[18]  Chin-Teng Lin,et al.  Noisy speech processing by recurrently adaptive fuzzy filters , 2001, IEEE Trans. Fuzzy Syst..

[19]  Bernard Widrow,et al.  Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[20]  Satoshi Nakamura,et al.  Robust speech recognition in car environments , 1998, Proceedings of the 1998 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP '98 (Cat. No.98CH36181).

[21]  Lotfi A. Zadeh,et al.  Fuzzy Logic Toolbox User''''s Guide , 1995 .