Wavelet-based speech enhancement : a statistical approach

[摘要] ENGLISH ABSTRACT: Speech enhancement is the process of removing background noise from speech signals. Theequivalent process for images is known as image denoising. While the Fourier transform iswidely used for speech enhancement, image denoising typically uses the wavelet transform.Research on wavelet-based speech enhancement has only recently emerged, yet it showspromising results compared to Fourier-based methods. This research is enhanced by theavailability of new wavelet denoising algorithms based on the statistical modelling ofwavelet coefficients, such as the hidden Markov tree.The aim of this research project is to investigate wavelet-based speech enhancement froma statistical perspective. Current Fourier-based speech enhancement and its evaluationprocess are described, and a framework is created for wavelet-based speech enhancement.Several wavelet denoising algorithms are investigated, and it is found that the algorithmsbased on the statistical properties of speech in the wavelet domain outperform the classicaland more heuristic denoising techniques. The choice of wavelet influences the quality of theenhanced speech and the effect of this choice is therefore examined. The introduction of anoise floor parameter also improves the perceptual quality of the wavelet-based enhancedspeech, by masking annoying residual artifacts. The performance of wavelet-based speechenhancement is similar to that of the more widely used Fourier methods at low noiselevels, with a slight difference in the residual artifact. At high noise levels, however, theFourier methods are superior.

[发布机构] Stellenbosch University