Speech Denoising with Maximal Overlap Discrete Wavelet Transform

In this paper, the effectiveness of the maximum overlapping discrete wavelet transform (MODWT) method on denoising the speech signal is tested and examined. Ensuring the intelligibility of the speech signal in noisy environments by separating it from the noise is a widely researched topic today. On the other hand, being able to recover the original speech from the noisy signal with minimal distortion is a challenge due to the difficulties in removing the background noise. Numerous factors in environmental noise environments can interfere with the signal. In this study, the performance of some discrete wavelets transform methods is experimentally analyzed using different wavelet filters. The analysis program was carried out in the MATLAB environment. As the input noise speech signal, speech sounds containing different environmental background noises (train, car, station, plane, etc.) were analyzed. During the tests, these noisy input signals were filtered out from the speech signal by wavelet analysis. The input noisy speech signal is decomposed into wavelet coefficients with different thresholding methods. The reconstructed speech was compared by measuring the signal-to-noise ratio (SNR) values between the noisy input signal and the smoothed output signals. The scientific contributions of the study include a detailed comparative analysis of the performances of various wavelet methods against different background environmental noises.