Video Subtool

Add subtitles to video by feeding audio to vosk and resulting text into (soft-)subtitled mkv. All spoken words are extracted into a json file for further processing, like searching in a subtitled video.

Requirements

Packages

• ffmpeg-python
• vosk

Download e.g. german from and rename folder to model.

Generating Subtitles / Extracting Words

subtool.py <input_video_filename>

Outputs:

• <input_video_filename>.mkv - Video with subtitles
• <input_video_filename>.json - List of sspoken words with timestamp and accuracy of detected words (directly from vosk)
• <input_video_filename>.ass - Subtitles in SubStation Alpha Format
• output.wav - Temporary Audio file