How do I convert a hdmv_pgs_subtitle (which is image based) to a text based subtitle in a MKV file?
I have tried ffmpeg -i in.mkv -c:v copy -c:a copy -c:s mov_text out.mkv but that results with the following error:
Stream mapping:
Stream #0:0 -> #0:0 (copy)
Stream #0:1 -> #0:1 (copy)
Stream #0:2 -> #0:2 (hdmv_pgs_subtitle (pgssub) -> mov_text (native))
Error while opening encoder for output stream #0:2 - maybe incorrect > parameters such as bit_rate, rate, width or height
Converting image based subtitles to text is a nontrivial process, as you will need some kind of OCR system to interpret the bitmaps and figure out what the corresponding text is. ffmpeg alone will not do that for you.
I am not aware of any app that will do the whole process in one go, for Linux/UNIX. However, this process should work:
mkvextract or ffmpegmkvmerge or ffmpeg
bdnxml2srt has vanished from the Internet. I've updated the answer with a different process (going through the DVD subtitle format instead) that should work. All the tools for that one are available.
This also seems to work as an online tool, but you have to wait a bit if there is a queue.
To extract the subtitles file, you can use this:
ffmpeg -i input.mkv -c copy -map 0:2 sub.sup
