A New Speech Enhancement Method Based On Wavelet Packet Transform

In this paper, a new speech enhancement algorithm based on the wavelet packet transform is proposed to suppress the noise. Before decomposing the noisy speech into componets by the wavelet space, we select and construct a new window function, which is smoother than the Hamming window in frequency domain to process the frame´s partition, in order to improve the wavelet thresholding performance in the wavelet domain, we define a new parameter to adapt in time the thresholds according to a nonlinear function of the wavelet coefficient energies. Experimental results show that the proposed approach improves the naturalness, and does not cause annoying residual noise in the enhanced speech. Visual inspection of spectrograms and listening experiments are used to support the results.