Speech Compression with Best Wavelet Packet Transform and SPIHT Algorithm

Abstract

This paper proposes a new coding algorithm of speech compressing, which is based on Best Wavelet Packet Transform and SPIHT coding algorithm. The new approach applies the wavelet packet transform to process speech signal to obtain optimal wavelet tree to allocate the dynamic bits, and then uses the modified SPIHT coding algorithm to compress the coefficients from the wavelet packet transform. The entropy coding is also used to increase the compression ratio. The experiments indicate that the new approach can gain better quality signal reconstruction at high compression ratios and have a low-complexity and low-delay.