基于ACF和AMDF的基音检测改进算法  

Improved Method of Speech Pitch Detection Based on ACF and AMDF

作　　者：李志军[1] 尹霞[1] 

机构地区：[1]扬州大学信息工程学院,江苏扬州225009

出　　处：《电声技术》2011年第1期50-52,55,共4页Audio Engineering

摘　　要：自相关函数法和平均幅度差函数法是基音周期检测的两种常用方法,两者具有独立且相似的统计特性。在讨论了传统基音周期检测算法及其缺点后,提出了综合应用ACF和AMDF函数进行基音检测的算法。该算法简单可行,性能优于一般传统时域基音周期检测算法。从仿真结果可以看出,对于含噪语音信号,该算法具有较强的抗噪能力。The Auto Correlation Funetion（ACF） and the Average Magnitude Difference Funetion（AMDF） are two commonly used method of pitch detection. They have independent and similar statistical characteristics. Based on the discussion of traditional pitch detection algorithm, a comprehensive application with ACF and AMDF is given. The method is simple and practicable. It＇s better than traditional time-domain pitch detection algorithm. Simulation results show that the algorithm has strong antinoise ability.

关 键 词：基音检测 自相关函数 平均幅度差函数 

分 类 号：TN912.3[电子电信—通信与信息系统]