Discrete Fourier Transform and Discrete Wavelet Packet Transform in speech denoising

The paper aims to emphasize the importance of computer simulation in speech denoising area by comparing Fourier Transform to Discrete Wavelet Packet Transform. Wavelet Packet Coefficients have the ability to capture important localized time and frequency features of speech. For denoising, certain Wavelet Packet Coefficients are set to zero and the others are thresholded. We show encouraging results using the denoised wavelet coefficients on small-scale experiments with speech corrupted by an artificial noise.