Title :
Optimal scalar quantization of the PARCOR coefficients for speech coding
Author :
Saoudi, S. ; Le Guyader, A. ; Boucher, J.M.
Author_Institution :
Dept. of Math. & Commun. Syst., ENST, Brest, France
Abstract :
The optimal scalar quantization of the PARCOR coefficients and of the log area ratios (LARs) for speech coding is treated. The kernel method to obtain the probability density function of the LARs, and the quantization tables for each coefficient is then computed with the Lloyd-Max algorithm. The optimum bit allocation is performed by the Fox-Makhoul procedure on the LARs and on the decorrelated LARs. The quantization tables of the LARs and the decorrelated LARs are given for 32 bits, using a code-excited linear predictive coder
Keywords :
encoding; probability; speech analysis and processing; Fox-Makhoul procedure; Lloyd-Max algorithm; PARCOR coefficients; code-excited linear predictive coder; kernel method; log area ratios; optimal scalar quantization; optimum bit allocation; probability density function; speech coding; Bit rate; Convergence; Decorrelation; Histograms; Kernel; Lattices; Probability density function; Quantization; Shape; Speech coding;
Conference_Titel :
Electrotechnics, 1988. Conference Proceedings on Area Communication, EUROCON 88., 8th European Conference on
Conference_Location :
Stockholm
DOI :
10.1109/EURCON.1988.11097
