基于多层次相似性检测机制的语音基频实时估计算法

摘要

To get the pitch of a speech remains an unsolved but basic problem in the field of speech signal process.The traditional method of estimating pitches based on autocorrelation function is not always accurate,while those which can give rather accurate results are complex and hard to apply to situations demanding good real-time performance.In order to take into account both the accuracy and the real-time performance, the authors have developed an algorithm based on a multi-level likeness checking mechanism to evaluate the pitch of speech signal.The algorithm has a good real-time behaviour and can be easily realized by hardware.%基音周期的确定是语音处理领域中的一个尚未完全解决的基本问题。传统的基于自相关函数的估计基音周期的方法过于粗糙，而那些准确性高的方法计算太复杂，不适用于实时要求高的场合。为了兼顾准确性与实时性，提出了一种基于多层次相似性检测机制的语音基频估计算法，该算法具有实时性好、易于硬件实现的特点。