摘要
提出了一种基于LPC谱和离散小波变换的改进自相关基音频率检测算法,用于检测病理嗓音的基音频率。首先采用LPC谱特性滤除高频,得到预测余量,然后采用6级db20离散小波变换处理预测余量信号,并结合自相关法提取基音频率。采用美国MEEI公司的病理嗓音数据库进行基频检测实验。实验结果表明,本方法能够有效的去除倍频、半频的影响,是一种有效的病理嗓音基音频率检测算法。
This paper proposes an improved pitch frequency detection algorithm based on LPC spectrum and discrete wavelet transform for detecting pitch frequency of the pathologic voice. Firstly, the high frequency is filtered out by LPC spectrum properties, thus getting the predicting residual signal. Then the level-6 db20 discrete wavelet transform is used to process the predicting residual signal, and in combination of the auto- correlation the pitch frequency is extracted. Finally the experiment with pathological voice database of MEEI Company in the United States indicates that this algorithm can effectively remove the multi-frequency or half -frequency influence, and is an effective detection algorithm of pathological voice-oitch freauenev.
出处
《通信技术》
2013年第10期100-102,共3页
Communications Technology
基金
国家自然科学基金(批准号:61271359)
关键词
病理嗓音
基音频率
LPC谱
自相关法
离散小波
pathologic voice
pitch frequency
LPC spectrum
autocorrelation method
discrete wavelets