IMPROVED SPEECH QUALITY AND EFFICIENT VECTOR ION IN SELP
D.J. Krasinski, R. H. Ketchu · 1988
(Stochastically Excited Linear Prediction) algorithm of a two-stage vector quantization using an adaptive and a stochastic codebook is described. The adaptive quantization is similar to a closed loop long-term filter if the predictor delay is more than one frame ngth. The performance of the adaptive codebook procedure is improved by extending its codebook to include candidate vectors synthetic excitation displaying a high level of algorithm is introduced, which through ry of the error criterion significantly reduces the rt required for the search through the adaptive rithm can also be used for stochastic codebooks pping candidate vectors. It is shown that for s in which neighboring candidates overlap for the quantization performance is as high as for g fully independent candidate vectors.