Introduction - If you have any usage issues, please Google them yourself
Using the autocorrelation function method pitch period, the sound signal is sampled every 20ms frame is divided into a plurality of frames, and each frame sequence uates the autocorrelation function, taken the first maximum point of correlation function calculated the pitch period, and frame as the horizontal, the vertical axis pitch drawing and remove outliers, and finally take the value of the area that is part of a larger portion of the figure as voiced pitch.