摘要
研究了G.729编码流说话人识别算法,提出了一种使用固定码书增益作为判决参数的G.729静音帧判决方法。将这一方法应用于低复杂度的G.729编码流说话人识别,去除压缩语音数据中的静音帧从而提高识别率。
The algorithm of low-complexity automatic speaker recognition in the coded G.729 stream was researched and a method using the fixed codebook gain value to choose silent frames in G.729 coded speech was introduced. The method was applied to low-complexity automatic speaker recognition in the coded G.729 stream to discard the silent frames.
出处
《电子技术应用》
北大核心
2008年第1期97-100,103,共5页
Application of Electronic Technique
基金
国家高技术研究发展计划863基金项目(2006AA01Z146)专项经费资助
