一种基于三阶统计量的基音周期检测新方法  

A New Method of pitch Period Detection Based on Wavelet Transform and Third-order Statistics

作　　者：赵知劲[1] 吴杰[1] 

机构地区：[1]杭州电子工业学院

出　　处：《信号处理》2001年第2期99-103,共5页Journal of Signal Processing

基　　金：浙江省自然科学基金;电科院预研基金资助项目

摘　　要：本文利用小波变换和三阶统计量的基本特性，提出了一种小波变换与三阶统计量结合的基音周期检测新方法；在低信噪比的基音周期检测中，本方法可以获得比其它基音周期检测法更精确的结果；文中还分析了压扩因子和门限因子Ｋ的选择，实验表明，Ｋ在一定范围内取值时对基音周期检测结果影响不大。In the light of the properties of wavelet transform and higher-order statistics, a new pitch detection method with the combination of wavelet transform and third-order statistics is proposed. For the low signal-to noise ratio (SNR), the performance of the method is better than other methods which had been presented. The selections of the scaling factor and the threshold factor K are discussed. The simulation results show that the threshold factor K has less effect on the result of pitch detection.

关 键 词：小波变换 三阶统计量 基音周期检测 语音信号处理 语音识别 

分 类 号：TN912.3[电子电信—通信与信息系统]