基于语音增强技术的AMBE编码算法研究  

Study on AMBE Coding Algorithm Based on Speech Enhancement

作　　者：徐云乾[1] 王宽仁[1] 吴金涛[1] 

机构地区：[1]空军工程大学信息与导航学院,陕西西安710077

出　　处：《计算机仿真》2012年第11期420-423,共4页Computer Simulation

摘　　要：参数编码算法在语音通信中噪声不能避免且无法完全滤除掉,造成提取参数不准确。针对提取参数精确度低的问题,提出了一种基于语音增强技术的先进多带激励语音编码算法。对语音增强算法的原理进行了简要的描述,提出采用AMBE语音编码算法中参数的提取方法。进行仿真的结果表明,经语音增强技术处理后的合成语音质量明显好于未经语音增强技术处理的合成语音,表明改进的编码算法能有效抑制噪声、提取出更加精确的语音参数,合成语音效果更好。Aiming at the characteristics that the parameter codes algorithm can not extract parameters accurately in case that noises in the speech correspondence are inescapability and can not be able to filter entirely, the paper pro- posed an advanced multi-band excitation speech encode algorithm based on the speech enhancement technique. Then the paper described the priniplc of speech enhancement algorithm succinctly and introduce the extracting method of the algorithm parameter in detail. The chart of the simulating experiment displays that the synthetic voice disposed by the speech enhancement technique is obviously better than the synthetic voice which is no disposal, which indicates that the encode a｜gorithm based on spech enhancement technique in this paper can eliminate noise so efficiently and extract speech parameter so accurately that the result of synthetic voice is better.

关 键 词：语音增强 参数编码 拟合误差 周期平滑 语音合成 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TN912.32[自动化与计算机技术—计算机科学与技术]