一种高质量短时延的4kb/s的语音编码算法  

A high quality vocoder with short delay at 4 kb/s

作　　者：秦龙[1] 成立新[2] 

机构地区：[1]重庆通信学院,重庆400035 [2]南京通信工程学院,南京210016

出　　处：《声学学报》2002年第4期363-368,共6页Acta Acustica

摘　　要：４ｋｂ/ｓ低速率语音编码是近年来语音信号处理研究的重要课题,也是国际电信联盟电信委员会(ＩＴＵ_Ｔ)下一步标准化的目标。本文提出一种混合激励的编码算法,它对浊音采用谐波激励,对清音采用码激励,该编码能在短时延(分析帧长为１０ｍｓ,采样率为８ｋＨｚ)的情况下获得很好的合成语音质量。本文分析了该算法的原理,并给出了模拟结果。Speech coding at bit rate 4 kb/s is an important problem in the field of speech signal processing, it's also the next goal for ITU-T making standard. A speech coding algorithm implemented by mixed excitation is provided in the article, whose voice and unvoice are excited by harmonics and codebook respective, The vocoder can output good quality speech with short delay (the length of frame is 10 ms,sample rate is 8 kHz). The principle of the algorithm is analyzed in the article, a result of test is also provided.

关 键 词：语音信号处理 原理 谐波激励 码激励 语音质量 语音编码 算法 短时延 

分 类 号：TN912.3[电子电信—通信与信息系统]