CS-ACELP语音编码算法的优化及其在DSP上的实现

Optimization of CS-ACELP Voice Code Algorithm and Its Implementation on DSP

针对CS-ACELP语音编码算法编码复杂度较高、DSP实时实现比较困难的问题,提出了一种可降低CS-ACELP语音编码算法复杂度的优化方法,分析了CS-ACELP语音编码算法原理,详细介绍了优化的CS-ACELP语音编码算法从固定码本搜索上降低算法复杂度的实现,并给出了在16位定点DSP芯片TMS320VC5402上实现CS-ACELP语音编解码方案的硬件及软件设计。实验结果表明,优化的CS-ACELP语音编码算法降低了运算复杂度,提高了运行速度,重建的语音符合标准的编解码要求。

In view of the problem that CS-ACELP voice code algorithm has higher coding complexity and is difficult to realize real-time on DSP, the paper proposed an optimization method that could reduce complexity of CS-ACELP voice code algorithm. It analyzed principle of CS-ACELP voice code algorithm, introduced implementation that optimized CS-ACELP voice code algorithm reduced complexity in searching fixed codebook, and gave design of hardware and software of the scheme of Code and decode of CS-ACELP voice implemented on 16 bits fixed-point DSP chip TMS320VC5402. The experiment result showed that the optimized CS-ACELP voice code algorithm can reduce operation complexity and improve running speed, and reconstructed voice meets the standard demands of code and encode.