Safety-net pyramid VQ of LSF parameters for wideband speech codecs

Abstract

In this paper, we propose the predictive pyramid VQ (PPVQ) with a safety-net concept, to quantize LSF parameters of the wideband speech. For prediction of the LSF parameters, the 1st order auto-regressive (AR) predictor is used. To quantize the prediction error signal having a Laplacian distribution effectively, we use a pyramid VQ (PVQ). For encoding low-correlation frames, a memoryless PVQ is used. By combining the PPVQ and the memoryless PVQ, hereby denoted "safety-net pyramid VQ", the advantages of both quantization methods can be utilized. The average SD value of a safety-net pyramid VQ is 0.35 dB less than that of a predictive split VQ (PSVQ). The safety-net pyramid VQ also requires low complexity and does not need any memory for the codebook storage.