Find differences between an audio and its transcript

[CorentinJ]

Made a small pip installable library for when you want to compare a speech segment with the text it's expected to say. Example (audio as mp4 to allow an embed):

sphere.mp4

from transcription_diff.text_diff import transcription_diff, render_text_diff

diff = transcription_diff("You can go pretty far in life if you're a perfect sphere in a vacuum", "sphere.mp4")
print(render_text_diff(diff))

! Well
You can go pretty far in life
! when
+ if
you're a perfect sphere in a vacuum

Installable as pip install transcription-diff
Project repo