In this paper, a new method is proposed for improvement of speech which is distorted by acoustic noise. Acoustic noise reduction is done through a proposed post processed adapted cascaded median filter based on spectral subtraction technique. This method use two stages of filter, in which background noise is eliminated by first stage cascaded median filter and then output speech is post processed by second stage adaptive filter, to reduce musical and residual noise. Proposed post processing algorithm is compared to conventional single stage cascaded median filter based on subjective listening tests and perception evaluation of speech quality (PESQ) scores. Simulation is done in Matlab-15 and results show that enhanced speech generated by proposed algorithm has better quality than conventional cascaded median filter.

Speech Enhancement, Noise Estimation, Spectral Subtraction, Cascaded Median Filter, Musical Noise

[1] S.F. Boll, “Suppression of acoustic noise in speech using spectral subtraction”, IEEE Transaction Acoustic, Speech, Signal Process., vol.27, no. 2, pp. 113-120, 1979.
[2] M. Berouti, R. Schwartz, and J. Makhoul, “Enhancement of speech corrupted by acoustic noise”, In the Processing of IEEE International Conference of Acoustic, Speech and Signal Processing, pp. 208-211, 1979.
[3] S. Kamath and P. Loizou, “A multiband spectral subtraction method for enhancing speech”, In the Processing of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), Volume 4, pp IV-4164, 2002.
[4] Ekaterina Verteletskaya and Boris Simak, “Noise reduction based on modified spectral subtraction method”, IAENG International Journal of Computer Science, 38:1, IJCS_38_1_10, 2011.
[5] Martin, R, “Spectral subtraction based on minimum statistics”, In the Processing of Eur. Signal Process. Conf., pp. 1182-1185, 1994.
[6] V. Stahl, A. Fisher, and R. Bipus, “Quantile based noise estimation for spectral subtraction and wiener filtering”, In the Processing of ICASSP, pp. 1875-1878, 2000.
[7] S. Rangachari, and P. C. Loizou, “A noise-estimation algorithm for highly non-stationary environments”, Speech Communication., vol. 48, pp. 220-231, 2006.
[8] I. Cohen, “Noise spectrum estimation in adverse environments: improved minima controlled recursive averaging”, IEEE Trans. Speech Audio Process., vol. 11, no. 5, pp. 466-475, 2003.
[9] Santosh K. Waddi, Prem C. Pandey, and Nitya Tiwari, “Speech Enhancement Using Spectral Subtraction and Cascaded-Median Based Noise Estimation for Hearing Impaired Listeners”, NCC, Department of Electrical Engineering Indian Institute of Technology Bombay, pp 1 - 5 , 2013.
[10] Kun-Ching Wang, “Wavelet-based speech enhancement using time-frequency adaptation”, EURASIP Journal on Advances in Signal Processing, 2009.
[11] Y. Hu and P.C. Loizou, “Evaluation of objective quality measures for speech enhancement”, IEEE Transactions on Audio, Speech, and Language Processing, volume. 16, pp. 229-238, 2008.
[12] G. Doblinger, “Computationally efficient speech enhancement by spectral minima tracking in subbands”, In the Processing of IEEE International Conference on Eurospeech, vol 2, pp 1513–1516, 1995.
[13] H. Hirsch and C. Ehrlicher, “Noise estimation techniques for robust speech recognition”, In the Processing of IEEE International Conference of IEEE International Conf. on Acoustic, Speech Signal Process., pp 153–156, 1995.
[14] C. Ris and S. Dupont, “Assessing local noise level estimation methods: application to noise robust ASR”, IEEE Transaction on Speech Comm. Vol 34, pp 141–158, 2001.
[15] B. Widrow and S. D. Stream, Adaptive Signal processing, New York: Prentice-Hall, 1985.