摘要
本文提出了一种有效而实用的基音提取算法-峰谷点修正自相关法。通过理论和实验研究,结果表明该方法克服了通常所采用的并联处理法抗噪声性能不强和自相关法不易实时实现的不足,而同时具有并联处理法容易实时实现及自相关法抗噪声性能强的优点。而且该方法简单、可靠、估计精确。
In this paper,we present an efficient and useful algorithm for pitch period extractioncalled peak -valley modified autocorrelation method. We make researches and deductionfrom theory and do some experiments with a computer,the results show that the proposedmethod overcomes disadvantages of parallel processing method and autocorrelationmethod. For parallel pr0cessing method can not obtain good results from a noise-comtaminated speech and autocorrelation method isn't generally easy to realize real-time implementation. And peak -valley modified autocorrelation method easily realizes real - timeimplementation and can obtain good results from a speech contaminated by noise. It hastheir advantages. Meanwhile the proposed method is simple and reliable and can obtain accurate results.
出处
《杭州电子科技大学学报(自然科学版)》
1996年第3期38-42,共5页
Journal of Hangzhou Dianzi University:Natural Sciences
关键词
语音信号
基音提取
自相关法
speech signal, pitch extraction, autocorrelation method