基于时长调整技术的低速率语音编码算法  

Low-bitrate Speech Coding Algorithm Using Time-scale Modification

作　　者：谢贵武 丁竑[1] 汤云革[1] 张雄伟[2] 杨继红[1] 

机构地区：[1]中国人民解放军63880部队 [2]解放军理工大学指挥自动化学院训练部

出　　处：《军事通信技术》2010年第1期51-55,共5页Journal of Military Communications Technology

摘　　要：研究并实现了一种新的低速率语音编码方案。通过将时长调整算法作为声码器的预处理和后处理,不需要修改声码器结构就能灵活地实现低速率语音编码。文中选择基于STFT的时长调整算法、SOLA算法、WSOLA算法分别与2.4kb/s和600b/s的声码器相结合,仿真结果显示,新方案进一步降低了语音编码速率,为低速率语音编码研究提供了一条有效的新思路。A new low-bitrate speech coding scheme was studied and implemented. Through appending TSM as previous and post process of the vocoder, low bit rates could be easily offered without modifying the vocoder structure. In this paper, the TSM based on STFT, SOLA and WSOLA were combined with 2.4 kb/s vocoder and 600 b/s vocoder separately. Simulation results indicate that the new scheme greatly decreases the speech coding rate and provides a novel idea for researching on low-bitrate speech coding.

关 键 词：低速率语音编码 时长调整 短时傅立叶变换 

分 类 号：TN91[电子电信—通信与信息系统]