基于正弦激励的3.6Kb/s低复杂度语音编码算法  

Low-Complexity 3.6 Kb/s Speech Coding Algorithm Based on Sinusoidal Excitation

作　　者：盛玉霞[1] 崔慧娟[1] 唐昆[1] 

机构地区：[1]清华大学电子工程系微波与数字通信国家重点实验室,北京100084

出　　处：《电声技术》2008年第1期56-59,共4页Audio Engineering

摘　　要：基于正弦激励线性预测技术,提出了一种3.6Kb/s的低复杂度语音编码算法。算法通过分子帧参数提取和多带混合激励技术,保证了语音质量;通过线谱对参数的全局最佳标量量化以及正弦波幅度的全1模型表示,降低了算法的运算量和存储量。测试结果显示,语音质量优于欧洲集群通信标准TETRA中的ACELP4.567Kb/s低复杂度语音编码器,而运算量和存储量不足ACELP的1/4。Based on sinusoidal excitation linear prediction method, a 3.6 Kb/s low-complexity speech coding algorithm is proposed. Sub-frame parameter extraction and multiband mixed excitation are used to enhance the performance of the algorithm. Meanwhile, the global optimum scalar quantizer is employed for the LSP parameters, and All-1 model is used to represent the sine wave＇s amplitude, thus the calculation and storage are decreased. Simulation results show that the proposed speech coder outperforms the TETRA-ACELP 4.567 Kb/s coder in speech quality and complexity.

关 键 词：正弦激励 语音编码 低复杂度 

分 类 号：TN912.32[电子电信—通信与信息系统]