Exploring dynamic time warping technique to wavelet speech compression

Wavelet analysis is a significant tool that has been used for speech and image processing. Three wavelet families, Haar wavelet (Haar), Biorthogonal wavelet (Bior) and Discrete approximation of Meyer wavelet (Dmey) are used in this experiment. Eighty speech signal, which are selected from male and female, are used as input signal. Dynamic Time Warping (DTW) technique is applied to determine and adjust speech feature for compression. The results shown that the Haar wavelet provides good performance in compression condition and also keep the quality of speech close to the original speech.