Design and performance of trellis vector quantizers for speech signals

Design algorithms are presented for linear predictive vector quantizers with trellis structures that, when combined with appropriate search procedures, can achieve lower distortions than conventional memoryless vector quantizers. Two types of trellis structures, shift register and minimum degradation network, are considered. In the shift register case (SR-TVQ), the state of the trellis encoder corresponds to the content of the shift register; it therefore does not require an explicit state-transition matrix. The minimum degradation network is a state-transition matrix obtained through a pruning procedure (P-TVQ) to give minimum distortion degradation for an omnisearch (full-rate) vector quantizer. The role of the search procedure, the delay requirements of each type of encoder, and the performance of the designs are discussed and the two methods are compared. >