I make short videos (5-10 minutes each) and have the script that actors used. I would like to combine both to make subtitles in the easiest way possible. The videos are on my hard drive and on my hidden Vimeo account. The language is European Portuguese.
One solution would be to detect silences in the audio track with thresholds of duration and volume, extract each non-silent portion into an audio file, transcribe it with speech-to-text, generate a subtitles file like VTT and adjust the thresholds. When the automatic subtitles line up well with the video, I could copy-paste from the original script to this subtitles file. This might take many hours in Python and some speech recognition system, then it would work for all other videos.
Has anyone found an easier or ready-made solution for this problem?