G.723.1语音编解码器优化策略的研究

Optimization of G.723.1 Speech Codec

在语音编解码器的实际应用中 ,语音编解码器的编解码时间是一个关键的性能指标。本文通过对 G.72 3.1语音编解码器客观测试和使用 VC+ +的工具 profile对 G.72 3.1语音编解码器测试结果的分析 ,提出对 G.72 3.1语音编解码器的优化应着重于对计算复杂度较大的模块进行优化 ,如基音估计、自适应码本搜索、固定码本搜索等模块。并针对这些模块给出相应的优化方法。最后对根据优化策略优化的代码和未优化代码进行了测试 ,结果显示优化代码比未优化代码的运行时间减少了 2 4 %～ 31%。

The running time of a speech codec is a key performance target. According to the analysis of the objective test requirements of G.723.1 codec, and the test results using profile tool in VC ++ for the codec, this paper first points out that the optimization of it should focus on the modules with large computational complexity, such as the pitch estimation module, and the adaptive and the fixed code book search modules. Secondly, the optimization methods for each module are given. Finally, the optimized and the original source codes are tested. Experimental results show that the running time of the optimized source code is decreased by 24%～31% of that of the original source code.