AMR语音编解码算法分析及优化  

Analysis and optimization of AMR Speech Codec

作　　者：黄桃[1] 李小文[1] 袁李林[1] 

机构地区：[1]重庆邮电大学,重庆400065

出　　处：《通信技术》2007年第12期391-393,共3页Communications Technology

基　　金：国家"863"计划引导项目(2004AA001390)

摘　　要：自适应多速率语音编解码技术共有八种速率,每一种速率都是基于代数码本线性预测技术语音编解码算法的。文中参照3GPPTS协议,分析了代数码本线性预测技术的编解码原理,并对其中的关键技术开环基音分析和固定码本搜索提出了优化方法。经过实验验证,改进后的算法降低了运算复杂度,并能获得良好的语音音质。AMR speech codec has eight rates, each of which is based on ACELP. This paper refers to 3GPP TS, analyzes the principle of ACELP speech codec, and presents new methods for optimizing its key technology, such as open-loop pitch analysis and algebraic codebook search. Experiment has proved that the computational efficient has been improved, and good speech quality achieved.

关 键 词：自适应多速率 代数码本线性预测技术 码本搜索 

分 类 号：TN912.3[电子电信—通信与信息系统]