[MPlayer-users] Dumping DVD subtitles to MPsub format

[Mike Albon]

On Fri, 2002-06-07 at 00:10, Arpi wrote:
> [Automatic answer: RTFM (read DOCS, FAQ), also read DOCS/bugreports.html]
> Hi,
> 
> > > > Is there a way to convert DVD subtitles (the ones I get using -sid)
> > > > into the funky MPsub format?
> > > No, dvd subtitles are graphics. It would need some kind of ocr. You must u
> > se
> > > some other tool to convert the dvd subtitles in any text subtitles format 
> > know
> > > by mplayer. Then you can convert them to MPsub.
> > > 	Albeu
> > What format graphics are they?  Maybe when I get a chance I can come up with
> >  something you can use...
> 
> rle-compressed 4-color (using yuv palette) bitmaps
> 
> writting an ocr for dvd subs isn't so hard at first look.
> as they are computer generated - the chars are pixel-by-pixel identical,
> so you have to teach teh ocr only once to each char and then it's go...
> problems:
> - it needs an interactive gui, for teaching the code to the chars
> - many problems appear when you have bold or italic text, so separating
> chars from eachother is no more a simple job :(

I was looking at hacking something together with Python using libgocr
(jocr.sf.net), but I couldn't get the graphics out. (I don't do C). Is
it too much to ask for an MPsub like script file pointing to graphics
files. Like 001.png 002.png etc.

Thanks

Mike