[MPlayer-users] extracting subtitles

[Raphael]

Rich Felker wrote:
> On Sat, Mar 25, 2006 at 06:56:33PM +0000, Raphael wrote:
>   
>> RC wrote:
>>     
>>> On Sat, 25 Mar 2006 17:20:36 +0000
>>> Raphael <mencoder at lesshaste.plus.com> wrote:
>>>
>>>   
>>>       
>>>> "[open] number of subtitles on disk: 0" which makes no sense as the
>>>> film is in French with subtitles.
>>>>     
>>>>         
>>> Are the subtitles hard-coded perhaps?  ie. Do they show up when you play
>>> the DVD with no -sid or -slang options?
>>>   
>>>       
>> Sadly I think they might be. Yes they do show up and they even stay
>> there after I disable subtitles by pressing "v". What is weird is that
>> ALL my DVDs appear to have hard coded subtitles (which is what made me
>> suspicious that the fault is with mplayer)!
>>     
>
> It should probably make you suspicious that your DVDs are bootleg.. :)
>
>   

But they are not!!   They are either bought from a proper shop, rented
or come free with the weekend paper :)
>> I don't suppose there is a tool for extracting hard coded subtitles?
>> Technically it should be easy as they sit in the black border below the
>> video...
>>     
>
> Ah, you're very lucky. Normally they're on top of the picture which
> really ruins it. Try this: use the crop filter to crop out the movie
> and show just the subtitles. Then use the decimate filter to remove
> duplicate frames. Then use -vo png to save the subtitles to files, and
> an OCR program to convert the results to text. You can probably come
> up with a way to automate getting the timestamps of the subtitles
> too..
>
> Rich
>   

Thanks. I see a mini project coming on :)

Raphael