Pitch detection based on autocorrelation square function and wavelet transform
基于自相关平方函数与小波变换的基音检测

A robust pitch detection algorithm based on autocorrelation square function and wavelet transform was proposed for speech signals severely corrupted by noise. Noisy speech signal was first preprocessed by wavelet transform. The autocorrelation square function had been applied to emphasize the peak of the true pitch period. This algorithm can get the exact pitch of the speech signals in strong noisy environment. Experimental results indicate that, compared with classical pure autocorrelation method, the proposed algorithm has better robustness and higher accuracy with lower computational complexity, which is good for speech synthesizing and coding in real-time.