data set is at Link: https://pan.baidu.com/s/1dlzGkABaMFH3e_qWJtNW9g passwd: xp8x
A simple speech recognition system which can is aimed at single word.
- find end point from the original sound file(zcr.py && volume.py)
- get mfcc feature from the processed sound file( framenum * 39)
- train model with SVM and mfcc data
- use reliable model to recognise new sounds