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。 There are eight rates for adaptive multi-rate speech codec, each of which is based on algebraic codebook linear predictive speech codec algorithm. With reference to the 3GPP TS protocol, the codec principle of algebraic codebook linear prediction is analyzed and the optimization method is proposed for the key technologies such as open-loop pitch analysis and fixed codebook search. After experimental verification, the improved algorithm reduces computational complexity and can obtain good voice quality.