Direct Recovery of Clean Speech Using a Hybrid Noise Suppression Algorithm for Robust Speech Recognition System

A new log-power domain feature enhancement algorithm named NLPS is developed. It consists of two parts, direct solution of nonlinear system model and log-power subtraction. In contrast to other methods, the proposed algorithm does not need prior speech/noise statistical model. Instead, it works by direct solution of the nonlinear function derived from the speech recognition system. Separate steps are utilized to refine the accuracy of estimated cepstrum by log-power subtraction, which is the second part of the proposed algorithm. The proposed algorithm manages to solve the speech probability distribution function (PDF) discontinuity problem caused by traditional spectral subtraction series algorithms. The effectiveness of the proposed filter is extensively compared using the standard database, AURORA2. The results show that significant improvement can be achieved by incorporating the proposed algorithm. The proposed algorithm reaches a recognition rate of over 86% for noisy speech (average from SNR 0 dB to 20 dB), which means a 48% error reduction over the baseline Mel-frequency Cepstral Coefficient (MFCC) system.

[1]  Rainer Martin,et al.  Noise power spectral density estimation based on optimal smoothing and minimum statistics , 2001, IEEE Trans. Speech Audio Process..

[2]  D. Howard,et al.  Speech and audio signal processing: processing and perception of speech and music [Book Review] , 2000 .

[3]  Yifan Gong,et al.  Robust Speech Recognition Using a Cepstral Minimum-Mean-Square-Error-Motivated Noise Suppressor , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[4]  Masakiyo Fujimoto,et al.  Noise robust voice activity detection based on periodic to aperiodic component ratio , 2010, Speech Commun..

[5]  Ephraim.  Speech enhancement using a minimum mean square error short-time spectral amplitude estimator , 1984 .

[6]  Li Deng,et al.  Estimating cepstrum of speech under the presence of noise using a joint prior of static and dynamic features , 2004, IEEE Transactions on Speech and Audio Processing.

[7]  Li Deng,et al.  Evaluation of the SPLICE algorithm on the Aurora2 database , 2001, INTERSPEECH.

[8]  Jeff A. Bilmes,et al.  MVA Processing of Speech Features , 2007, IEEE Transactions on Audio, Speech, and Language Processing.

[9]  S. Boll,et al.  Suppression of acoustic noise in speech using spectral subtraction , 1979 .

[10]  Alejandro Acero,et al.  Acoustical and environmental robustness in automatic speech recognition , 1991 .

[11]  David Pearce,et al.  The aurora experimental framework for the performance evaluation of speech recognition systems under noisy conditions , 2000, INTERSPEECH.

[12]  Jacob Benesty,et al.  New insights into the noise reduction Wiener filter , 2006, IEEE Transactions on Audio, Speech, and Language Processing.

[13]  R. G. Leonard,et al.  A database for speaker-independent digit recognition , 1984, ICASSP.

[14]  Richard J. Povinelli,et al.  Minimum Mean-Squared Error Estimation of Mel-Frequency Cepstral Coefficients Using a Novel Distortion Model , 2008, IEEE Transactions on Audio, Speech, and Language Processing.

[15]  Rainer Martin,et al.  Spectral Subtraction Based on Minimum Statistics , 2001 .