The effects of vector transform on speech compression

Abstract

Until now, many techniques have been developed for speech compression. In this study, firstly, linear vector quantization is used to compress speech signals. Then, Linde-Buzo-Gray (LBG) algorithm, which is used for image compression, is adapted to speech signals for compressing process. Before this process, vector transform (VT) that is defined at second chapter is applied on speech signals. After the VT, speech vectors are coded using vectoralized LBG algorithm. Inverse VT is applied to decoded data. Obtained compression results are evaluated using graphics and SNR values. © 2007 Springer.