Speech Enhancement using Statistical Estimators Based on Wavelet Transformations  
  Authors : V. Vijaya Lakshmi; Dr. V. V. K. D. V Prasad

 

Estimators for speech enhancement by using wavelet transform is the new technique which is proposed in this paper. Here, we proposed a new set of estimators called magnitude square spectrum estimators beyond the conventional magnitude, power estimators using wavelet transform. Maximum a posteriori(MAP), Minimum Mean Square Error(MMSE) Estimators are derived using hard masking then Soft Masking by Incorporating a Posterior SNR uncertainty (SMPO),Soft Masking by Incorporating a Priori SNR uncertainty(SMPR) Estimators are derived using soft masking with wavelet transformations. These estimators are evaluated using the parameters Mean Square Error (MSE), Perceptual Evaluation of Speech quality (PESQ), Signal to Noise Ratio(SNR). The results showed that these magnitude square spectrum estimators reduces the distortion because of the localization property of the wavelet transform and increases the quality and signal strength of the speech signals.

 

Published In : IJCAT Journal Volume 2, Issue 7

Date of Publication : July 2015

Pages : 219 - 226

Figures :06

Tables : 06

Publication Link :Speech Enhancement using Statistical Estimators Based on Wavelet Transformations

 

 

 

V.VIJAYA LAKSHMI : received the B.Tech degree in electronics and communication engineering in 2013 from Amrita Sai Institute of Science and Technology, M.tech pursuing Digital Electronics and Communication Systems in Gudlavalleru engineering college.

Dr. V.V.K.D.V. Prasad : presently working Professor of ECE in Gudlavalleru engineering college from 2013 to till date. Current area of research in denoising of signals and images using wavelets transforms and S transforms.

 

 

 

 

 

 

 

Estimator

speech enhancement

MAP

MMSE

Soft Masking

SMPO

SMPR

From the results, it is proved that the magnitude square spectum estimators with wavelet transforms increases the Speech quality(PESQ),signal to noise ratio(SNR) and decreases the Mean square error(MSE).Out of all the five estimators derived it is observed that SMPO(soft masking incorporating posteriori SNR) shows the best performance.

 

 

 

 

 

 

 

 

 

[1] Gopalakrishna, V., Kehtarnavaz, V., Mirzahasanloo, T. and Loizou, P. (2012). "Real-time automatic tuning of noise suppression algorithms for cochlear implant applications," IEEE Trans. Biomedical Engineering, [2] Chen, F. and Loizou, P. (2012). "Contributions of cochlea-scaled entropy and consonant-vowel boundaries to prediction of speech intelligibility in noise," Journal of the Acoustical Society of America,131(5), 4104– 4113. [3] ”Noise estimation Algorithms for Speech Enhancement in highly non-stationary Environments”by Anuradha R. Fukane, Shashikant L. Sahare, IJCSI International Journal of Computer Science Issues, Vol. 8, Issue 2, March 2011. [4] Evaluation of Objective Measures for Speech Enhancement Yi Hu and Philipos C. Loizou Department of Electrical Engineering University of Texas at Dallas Richardson, TX, USA. [5] Erkelens J, Jensen J, Heusdens R. A data-driven approach to optimizing spectral speech enhancement methods for various error criteria. Speech Communication. 2007;vol. 49(no. 7–8):530–541. [6] Cohen I. Relaxed statistical model for speech enhancement and a priori SNR estimation. IEEE Trans. Speech and Audio Processing. 2005 Sept;vol. 13(no. 5):870–881. [7] Wolfe PJ, Godsill SJ. Efficient alternatives to Ephraim and Malah suppression rule for audio signal enhancement. EURASIP Journal on Applied Signal Processing. 2003;vol. 2003(no. 10):1043–105.