Jakobovski / free-spoken-digit-dataset

A free audio dataset of spoken digits. An audio version of MNIST.
626 stars 248 forks source link

.wav encoding for speaker Nicolas not consistent with other speakers #17

Closed slegroux closed 6 years ago

slegroux commented 6 years ago

FYI file encoding for speaker nicolas are 8bit unsigned integer whereas all other speakers are 16bit Signed int sox -b 16 -e signed-int old_nicolas.wav new_nicolas.wav does the trick

cag472 commented 6 years ago

Nicolas also seems to be stereo (2 channels) instead of mono. This seems to solve both problems: ffmpeg -i data/2_nicolas_4.wav -af 'pan=mono|c0=c1' nicolas_mono/2_nicolas_4.wav

Jakobovski commented 6 years ago

Fixed