View Full Version : Trying to OCR idx/sub: normal routes failing!


cultural_sublimation
13th November 2007, 14:56
I have subtitles in the idx/sub format pair. These are the only files I have. I am trying to OCR them into srt, but I've run across a number of problems:

a) the latest SubRip only seems to recognise a few of the pictures. Most of the bitmaps appear as a jumbled mess.

b) Subresynch recognises all the bitmaps without any problems. However, it doesn't seem able to properly OCR the text: it shows up whole words as single unit, instead of individual letters.

Since the new Subrip has the option to OCR subtitles in AVI files, I thought I had a clever idea: simply render the idx/sub into a "black" film. However, it seems that mencoder is unable to burn idx/sub subtitles.

Anyway, I'm kind of stuck now. Does anyone know of any other software to either a) OCR subtitles, or b) Burn subtitles into a "blank" avi?

Thanks for your help!
C.S.

unskinnyboy
13th November 2007, 15:03
a) the latest SubRip only seems to recognise a few of the pictures. Most of the bitmaps appear as a jumbled mess.Can you post a screenshot of this 'jumbled mess'?

manusse
14th November 2007, 20:53
Hi,

You can to use SubtitleCreator's OCR that relies on Office 2003 if you have it.
Otherwise you can also use SubtitleCreator to convert your sub/idx files to a sup file then use DVDSubEdit to OCR the subtitle and get a srt file.

Cheers
Manusse

cultural_sublimation
21st November 2007, 16:19
Thanks for all your replies! It worked: I used Subtosup (a small utility linked from the Doom9 downloads page) to convert the idx/sub to a sup file. Then I used DVDSubEdit to OCR the sup into a srt.

I wouldn't say it's practical (these are Windows applications, so you need to run them inside a virtualised environment or under Wine), but it does get the job done.

Cheers,
C.S.