Easy (?) subtitle problem ?

[Bende]

Hey guys,

usually i solve my technical problems via google but I can't find anything fitting my problem. My girlfriend hasn't seen me out of my room for hours now and i start feeling frustrated. So I registered ...

I have a original DVD (Avatar) with a "french subtitles"-option on it. All I want to do is write down the vocabulary from the subtitles. Then learn it in a couple of days and then watch the movie with her. And I do not want to see the movie actually and pause it all the time.

So my plan was to extract the subtitles to an .srt file. But that doesn't work. I have successfully extracted .sub and .idx from the DVD with the program AVIDEMUX (I'm running on Ubuntu).

But when I try to get the .srt file via "Tools / OCR (VobSub -> srt)" I fail completely for 2 reasons:
1) At the OCR recognition part I'm asked to fill in the letters of the glyph text. But it's very often that this glyph text is nothing but a graphic error !? There are no characters in the glyph text, it looks like ants-war !
2) The glyph text is in ENGLISH despite the fact that I have picked FRENCH when I selected the .idx file ... !?

I would be happy for a solution to EITHER get the .srt file OR get some kind of image that only shows the text of the subtitles (Which is enough for my purpose). I think .srt files are interesting for people who want to use them for another movie file. But I really only want to SEE/VIEW the subtitles so that I can pick the words that I don't know and learn them - so maybe there is a kind of graphical solution by substracting images?

Edit: There is a tutorial for the windows-program subrip: http://www.doom9.org/index.html?/mpg/dvdsubs.htm
As far as i understand, it saves the subtitles as many bitmaps. That's exactly what I need. So I've installed wine and tried to run subrip with it. And what I get is this:

I start subrip and click on "VOB". Then i click on "Open IFO" and browse to the DVD. There are several IFO's to pick, I choose (?): "VTS_05_0.IFO"



He loads some files and a list appears. The first file is unchecked -> do I need to check mark it?
"French" appears multiple times: Why and which to choose?
I check mark "Save Pictures as BMP"



After clicking on "Start" he wants me to enter some filename. I choose "fra" and click "Save":



Then the program starts and freezes at this position:



So what's the problem?

[MilesAhead]

Since you have Wine installed I would try to run DVDSubEdit. It has a built in OCR that I've used with good results. At least it works for me a high percentage of the time for English subs. If the French subs have lots of accented characters then it may not work so well. It's a good program to have in the toolbox. If you get it working, even if it doesn't work in this case, the effort won't be for nothing. I can't stand to do OCR. I hear people say they type a few characters and then it's all good. Never works that way for me. If automatic doesn't work, then I find another way. Like searching for the srt file online or something.

Also if you do get DVDSubEdit working, try to input ifo/sup rather than idx/sub. It has a better chance with sup I think. But I haven't used it hundreds of times. I'm relating my experiences.. which I admit are anecdotal as they say.

[Bende]

Since you have Wine installed I would try to run DVDSubEdit. It has a built in OCR that I've used with good results. At least it works for me a high percentage of the time for English subs. If the French subs have lots of accented characters then it may not work so well. It's a good program to have in the toolbox. If you get it working, even if it doesn't work in this case, the effort won't be for nothing. I can't stand to do OCR. I hear people say they type a few characters and then it's all good. Never works that way for me. If automatic doesn't work, then I find another way. Like searching for the srt file online or something.

Also if you do get DVDSubEdit working, try to input ifo/sup rather than idx/sub. It has a better chance with sup I think. But I haven't used it hundreds of times. I'm relating my experiences.. which I admit are anecdotal as they say.

Thank you for helping.
But I don't think OCR is a good idea. It seems to be prone to errors and only double the trouble (See my first post considering acidrip). That's why I wanted to get the bitmap-files with subrip which is my prior interest now - but I encounter the error described in the pictures.

But you made me rethink the idea of downloading a srt file. I had refused it because:
1) How to trust a source? I mean I'm a little panical about viruses ...
2) Aren't there many different srt files out there? I want to have exactly the text that is shown on my DVD... extracting the subtitles myself would fulfill this goal 100%
2) When I search for french subtitles for avatar, I always get the whole MOVIE with french subtitles and don't know which file to pick
3) Can you give me an example like where to find a safe download for avatar french srt?
4) Do you have some answers considering the subrip-pictures ?

[MilesAhead]

I don't want to use OCR either. See my first post. The point is, DVDSubEdit either works automatically, or it doesn't. You don't type in to tell it what the symbol is. You get .srt output. If there are errors that's another matter. I suggest search for the underscore character '_' all the bad lines I got had at least one of those. At least the ones I've done in English I got over 90% requiring no edit whatsoever.

There are good download sites in the stickies. I don't see the point of rehashing that info.

[Budman1]

I just ripped Transformers and used http://en.flossmanuals.net/avidemux/ch019_extract-dvd-subtitles/ as a reference in how to extract the IDX and SUB files from the VOB files using AVIDemux . The only difference is after you have them, use subtitle EDIT to drag the sub file into it and adjust the colors of the OCR to Black, Black, White, Black for readability.
. Y
Select your language and start OCR. It really does a good job with the white letters and outline and background set to Black. Be sure to click custom colors also.


BTW if you need languages added to subtitle Edit got to http://code.google.com/p/tesseract-ocr/downloads/list and download needed language and unzip the XXX.traineddata file (E.G. fra.traineddata) to Subtitle Edit_SE33\Tesseract\tessdata folder and it will be selectable within Subtitle Edit after that.