LD-CELP语音编码算法中矢量量化过程的改进  

An Improved Scheme of Vector Quantization in LD-CELP Speech Coding Algorithm

作　　者：黄德智[1] 马尽文[1] 

机构地区：[1]汕头大学数学研究所,广东汕头515063

出　　处：《电子学报》2001年第10期1415-1417,共3页Acta Electronica Sinica

基　　金：国家自然科学基金 (No .1 970 1 0 2 2 5);广东省自然科学基金 (No.970 377)

摘　　要：本文介绍了LD CELP算法的基本原理 ,在分析其编码的矢量量化过程的基础上提出了一种旨在提高编码速度的改进方案 .模拟实验表明 ,改进算法的编码速度平均提高了一倍 .虽然信噪比有所下降 ,但下降幅度仅为1 2dB 。In this paper,the basic principle of LD CELP (Low Delay Code Excited Linear Prediction) Algorithm is introduced.Based on a detailed analysis of the vector quantization of the algorithm,an improved scheme is proposed to increase the coding speed of LD CELP algorithm.It is shown by simulation experiments that with the improved scheme,the coding speed has been doubled,while the SNR has decreased by about 1 2dB which is small enough to guarantee the quality of the coding.

关 键 词：LD-CELP算法 矢量量化 语音编码 

分 类 号：TN912.3[电子电信—通信与信息系统]