语音LSF参数M-L搜索VQ算法研究  

M- L VQ Algorithm Research Based on the Speech LSF Parameters

作　　者：许晶晶 

机构地区：[1]广州广晟数码技术有限公司,广东广州510640

出　　处：《电声技术》2014年第6期53-56,71,共5页Audio Engineering

摘　　要：针对高效低速率语音编码,以LBG矢量量化码书设计算法为基础,研究了M-L搜索多级矢量量化(VQ)的码书设计算法和M-L搜索多级矢量量化编解码算法,同时对整个算法进行了全面的测试和性能分析。设计结果表明:该方法可有效提高LSF参数压缩的效率,改善谱失真指标。The M - L VQ code book design algorithm and coding algorithm focus on high efficiency and low bit rate speech coding are studied, which is based on the LBG - VQ code book design algorithm. The comprehensive testing and perform- ance analysis perform are also performed for this M - L VQ algorithm in this paper. Design result shows this algorithm can improve the compression efficiency of LSF parameters and improve spectral distortion indicator.

关 键 词：语音编码 线性预测 矢量量化 码书设计 

分 类 号：TN912[电子电信—通信与信息系统]