Why? Detecting whether there is speech or not should be enough to align video and subs. It's not like the word order would be scrambled in the subtitle file and needs to be sorted according to the video. :)