摘要： 为降低固定码本搜索算法的复杂度，在脉冲取代法的基础上提出一种码矢分段优化的快速搜索方法。采用码矢分段优化的方法，在保证语音质量的前提下，降低计算复杂度。实验结果表明，与AMR-WB采用的深度优先树算法及传统的脉冲取代算法相比，在不影响语音质量的条件下，码矢分段优化算法复杂度降低了70%~80%。

关键词: 语音编码, 固定码本搜索, 码矢分段优化, AMR-WB标准, MOS值

Abstract: In order to reduce complexity of fixed codebook search, a fast algebraic codebook search argorithm based on pulse replacement procedure is proposed in this paper. The proposed method is employed dividing candidate codevector into two sub-codevectors to optimize. It not only lowers the computational complexity sharply, but also ensures the speech quality. Experimental results show that the proposed method reduces the total computational complexity by 70%~80% with speech quality equivalent, compared with that by the depth-first tree search method adopted in AMR-WB and traditional pulse replace method.

Key words: speech coding, fixed codebook search, sub-codevector optimization, AMR-WB standard, MOS score

中图分类号: 

• TN912.3