Speech Enhancement using Maximal Overlap Discrete Wavelet Transform
Selma Özaydın, Iman Khalil Alak · DergiPark (Istanbul University) · 2018
Signal denoising for non-stationary digital signals can be effectivelysucceeded by using discrete wavelet transform. Selecting of a suitable thresholdingmethod is important to minimize the loss of useful signal information. Thispaper demonstrates the application of the maximal overlap wavelet transform(Modwt) technique in speech signal denoising. The analysis algorithm was performedon Matlab platform. In this algorithm, different kinds of input noisy speech signalsincluding environmental background noises such as restaurant, car, street orstation were tested. The noisy signals were filtered from the speech signal by thresholdingof wavelet coefficients with threshold estimation methods known as sgtwolog, modwtsqtwolog,heursure, rigrsure and minimaxi. The performance of the Modwt in denoisingprocess was evaluated by comparing signal-to noise ratio (SNR) and mean squareerror (MSE) results to those of well-known threshold estimation methods. First,denoising effectiveness of a Modwt based threshold method was tested indifferent scenarios and very important improvements in denoising process wereachieved by Modwt based scenarios. Next, the influence of the differentwavelets families on Modwt based threshold estimation method was evaluated by experimentalresults. The results revealed that Modwt based method outperforms conventional thresholdmethods while providing nearly up to a %24 increase in SNR value.