Creates an SRT file from a video file that has hardcoded subtitles. For example,
will generate
57
00:03:20,200 --> 00:03:22,120
就是那涌泉村的几户
The script relies on Tesseract for the optical character recognition.
- Python 3.9
- Tesseract OCR with support for your target language
- pipenv
To download the remaining python package dependencies, run
pipenv install
To extract hardcoded subtitles from a video file run:
pipenv run python hard_subs_to_srt.py "path/to/video_file.mkv" "path/to/subtitles.srt"
This opens a GUI where you can see the script working on the video. Commands that are available when running the video are:
- P - pause the subtitle extraction
- C - countinue the subtitle extraction if paused
- Q - stop the extraction and quit
For more information, have a look at the help documentation of the script by running:
pipenv run python hard_subs_to_srt.py -h
Note that the script is currently locked for specific format of video input. To get it working for your video you need to edit the script. So far I have only tried the script on 4k Chinese video. For lower resolution, the subtitle images may have to be scaled up before feeding them to Tesseract.