Extracting subtitles

[geraldholdsworth]

Hi all,

I've got a VOB file which plays in VLC and I can display subtitles (in a variety of languages). But, I want these subtitles, in the appropriate languages, in a text file with timecode markers. This is not from a commercial DVD, so these subtitles won't be online.

Anyone got any ideas?
BTW, I'm using a Mac.

Cheers,

Gerald.

[sP1d3r]

I want these subtitles, in the appropriate languages, in a text file with timecode markers.

Called a transcript, I believe.

[Boydie]

Ffmpeg? Don’t know if it preserves the timecodes, but there must be an option…

https://trac.ffmpeg.org/wiki/ExtractSubtitles

VLC can allegedly do it as well. Apparently Handbrake can too.

[geraldholdsworth]

Tried Handbrake - couldn't work out how to get it to do it.

Couldn't figure out how to install ffmpeg. Then I found out I can do it through homebrew. Still won't touch the subtitles. It claims that there isn't any...but there is. Apparently, 16 languages - only need the 6 of them (well, the client says we don't need Japanese, so that makes it 5).

[Boydie]

Is the VOB all you have, or do you have a DVD (either physical or iso)?
It seems not all subtitle data on DVDs is stored in the VOB; some is stored in other files. Players such as VLC seem to be able to cope without the missing data, but transcoders are apparently more fussy.

If you have the “DVD”, maybe try transcoding the relevant title(s) into individual container file(s) such as mkv (remembering to include the subtitles). Hopefully ffmpeg will have more luck in finding the subtitle data in the mkv than vob.

[Vivoii]

There are various tools for subtitle extraction but if you want them as a text file you’ll need optical character recognition (OCR) software as well.

The best tool I’ve used for both extraction and OCR is Subtitle Edit (freeware). It’s for Windows and Linux, but I’ve used it on a Mac, running within Crossover (an app that lets you run Windows software without installing a virtual machine).

There’s a guide here: https://iamscum.wordpress.com/guides/ocr/

[sP1d3r]

This relies on Windows programs to extract the subtitles but they'd subsequently need converting into a transcript which may not be possible;

https://www.youtube.com/watch?v=ru1aDajSe9g

[sP1d3r]

Perhaps converting into a Youtube video and then using Youtube transcription would work.

[scruss]

Couldn't figure out how to install ffmpeg. Then I found out I can do it through homebrew. Still won't touch the subtitles. It claims that there isn't any...but there is.

What does ffmpeg -i yourfile.vob say about subtitles? If it can't find subtitle text streams, then your client's video file has the subtitles "burned in" as image frames with no explicit text. Those require the OCR tricks described above, and lots and lots of proof-reading. This can be expensive and slow to get right.

[geraldholdsworth]

Is the VOB all you have, or do you have a DVD (either physical or iso)?

I've got a ripped copy of the DVD (so I've got the IFO and BUP files too in a VIDEO_TS folder. I can get hold of the physical DVD.

I've got my collegue working on it too - he's managed to extract the English subtitles, so far, into an SRT file, which I can use. Just need the other five languages.

[geraldholdsworth]

What does ffmpeg -i yourfile.vob say about subtitles? If it can't find subtitle text streams, then your client's video file has the subtitles "burned in" as image frames with no explicit text. Those require the OCR tricks described above, and lots and lots of proof-reading. This can be expensive and slow to get right.

It says:

Code: Select all

Input #0, mpeg, from 'Urquhart_Castle_Cinema.VOB':
  Duration: 00:08:38.02, start: 0.060000, bitrate: 8971 kb/s
  Stream #0:0[0x1bf]: Data: dvd_nav_packet
  Stream #0:1[0x1e0]: Video: mpeg2video (Main), yuv420p(tv, bottom first), 720x576 [SAR 16:15 DAR 4:3], 6000 kb/s, 25 fps, 25 tbr, 90k tbn
    Side data:
      cpb: bitrate max/min/avg: 6000000/0/0 buffer size: 1835008 vbv_delay: N/A
  Stream #0:2[0x85]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  Stream #0:3[0x84]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  Stream #0:4[0x83]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  Stream #0:5[0x82]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  Stream #0:6[0x81]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s
  Stream #0:7[0x80]: Audio: ac3, 48000 Hz, 5.1(side), fltp, 448 kb/s

I know that there are 6 (at least) audio streams, one for each language. VLC can display the subtitles, in the different languages.

[scruss]

No text subtitles found in your ffmpeg output, alas. Your colleague's SRT must've used the OCR method

[rmbrowngr]

I’ve used subtitle edit to ocr subtitles https://www.nikse.dk/subtitleedit. It’s free and quite useful.

[geraldholdsworth]

No text subtitles found in your ffmpeg output, alas. Your colleague's SRT must've used the OCR method

VLC greys out the Subtitle Track sub menu until the narrator starts speaking. Then I get a choice of about 16 tracks, and can change the font, colour, size, etc. which makes me believe they're in there, just not at the start.