Performance analysis of the speech enhancement application with wavelet transform domain adaptive filters

Abstract

Adaptive filters are one of the most commonly used methods in digital signal processing today. Nonetheless depending on the characteristics of the signals and noise the processing complexity and convergence speed for adaptive filters vary. The application of adaptive filters in the transform domain is preferred as a solution to this problem. It has been shown that the application of the NLMS (Normalized Least Mean Square) algorithm in the wavelet transform domain was successful for speech enhancement application. However further analysis is required to see the performance of the Wavelet Transform Domain (WTD)-NLMS method for cleaning speech signals disturbed by commonly used ambient noises for speech applications. Obtained results were evaluated with the measures frequently used for speech enhancement applications and compared with the results in the state-of-art. It was observed that the proposed WTD-NLMS structure outperforms speech enhancement applications done up to now in terms of SDR MSE STOI and PESQ metrics. © 2023 Elsevier B.V. All rights reserved.