Human Pitch Detection Program (Cepstrum and Autocorrelation algorithms)
This program is used to detect the pitch of human voice. Sound capture part of program is based on RTCP classes but pitch detection unit is famous Cepstrum algorithm.
In second version of program, autocorrelation algorithm is also added to the project for pitch detection.
You could use the radio button for algorithm selection.
autocorrelation algorithm could be used as a VAD (voice activity detection ) module.
This program could be used only in mono mode. It is better to set sampling frequency more than 16KHz and less than 40KHz.
Pitch Frequency is the base frequency of vocal cords. Pitch of male people are lower than of female ones.
To measure the pitch frequency,after setting sampling frequency and activating program by pushing "open device" button, make a continuous voice into microphone.
When you see a smooth vertical line on the chart press "Freeze" button.
You could use mouse to select pitch section of diagram. For that, drag mouse from upper left of desired area to lower right of it.
Then selected area will be zoomed and you could exactly measure pitch frequency.
For normal diagram show, drag mouse from lower right to the upper left of the chart.
Chart diagram of program is the project that was developed by Cedric Moonen and downloaded from Codeproject.
FFT library of program is KissFFT library written by Mark Borgerding.
MATLAB source code of the cepstrum algorithm is developed by Naotoshi Seo.
You could download executable 32 bit version of program using following link:
|rar package of human pitch detect program|
Copyright © 2013 DSPCottage, All rights reserved.