Ripping subs with OCR from DVD image

[carlmart]

I remember I used a program, many years ago, that read subtitles imprinted on the video image.

Can you refresh my mind on which programs do that?

It was far from perfect, and I never got to make it work completely.

But I would like to give it another try.

[Baldrick]

You can try subrip, File->Open hard subbed video files.

[devilcoelhodog]

And you can try another one too: AVISubDetector

This program is more complex than subrip, but it seems can do the job too !

PS: probably this program do not process VOB / DVD content directly. Before maybe you must to to convert your files into one single AVI ( using something like xvid / divx codecs) container, for example.


See the guide:


---> http://www.wxfs.org/guides/HowtoRiptheTimingandEnglishSubsFromanAVIfileUsingAviSubDetector.pdf


And more:


---> http://forum.doom9.org/showthread.php?t=121635


---> http://forum.doom9.org/showthread.php?t=75192


---> http://forum.doom9.org/showthread.php?t=144161


---> https://forum.videohelp.com/threads/296435-Rip-hard-coded-subs-from-avi-VirtualDub-filt...ubRip-problems


---> http://www.softpedia.com/get/Multimedia/Video/Other-VIDEO-Tools/PodPis.shtml



Best regards.


devil (johner)

[carlmart]

OK, I converted the DVD to sole mpeg file with PGCdemux, which I can open with Subrip.

But I don't know how to make it identify the characters and convert them into letters.

[devilcoelhodog]

Hello.

You can see this link below. Maybe it helps:


---> http://redonesubs.blogspot.ca/p/extracting-hardsubs.html


---> http://zuggy.wz.cz/guides/video.htm


Best regards.


devil (johner)

[guest]

Hi, I also tried to get some subtitles from my DVD, and following other users responses I got this method working for me. If you have any extra time you can give this a shot. Open PGCDemux, choose the mode/stream you want, and then under 'Options', choose 'Demux all subpic streams'. This will get you a .sup file. Open the .sup file in BDSup2Sub. Below the panel saying 'File/Edit/Settings/About', head to 'Output Format' and select 'Sub/IDX' and also from 'Palette' choose 'Create New'. Now go to Edit > Edit DVD Frame Pallete > and you must change the background colour to black! > After you have changed the colour, above the 'Ok' button select 'Set All' to set this colour scheme to all subtitles. Then press OK. (e.g. For my one the inside colour 1 is green, background colour 2 is grey... I changed colour 2 to black, colour 1 to white). Then export and save the sub/idx. You can now open Subtitle Edit, drag in the sub, and OCR from there (I use OCR via Image Compare, and play around with 'No. of pixels is space'). Hope this helps!

[devilcoelhodog]

a program, many years ago, that read subtitles imprinted on the video image.

If the subtitles are embedded, "burned", encoded on the video, it's not possible to separate or demux them, I think.

In this case, maybe OCR method can "read" the subtitles imprinted on the video image, using programs like AVISubDetector, Subrip or similar ones.

If the case is this, see the links mentioned before.

Now, if the subtitles are NOT embedded on the video and they exist as separate tracks, as we normally see on DVD's, Bluray's or even MKV files, you can try the tips suggested by @guest.

I point some other tips for such cases too:


---> http://www.afterdawn.com/guides/archive/rip_subtitles_with_subrip.cfm


---> http://en.flossmanuals.net/avidemux/ch019_extract-dvd-subtitles/


---> http://www.nikse.dk/SubtitleEdit/Help#importvobsub


Thanks.


devil (johner)

[manono]

Hi, I also tried to get some subtitles from my DVD, and following other users responses I got this method working for me. If you have any extra time you can give this a shot.

Not relevant to the OP's problem as his are burned into or embedded in the video image. There's no SUP file to demux.

[guest]

Not relevant to the OP's problem as his are burned into or embedded in the video image. There's no SUP file to demux.

My bad, the only experience I have with hard subs is also using SubRip as suggested above.

[carlmart]

Yes, unfortunately the subs are embedded on the video. The other way is easy and I do it all the time.

[carlmart]

Sorry, an alternate problem. It has to do with this though.

I'm having problems making AutoGK work. It stops when it gets to Virtual-Dub. It seems to do the processing but then it stalls there and doesn't finish the conversion. I wonder whhat is missing.

I had a similar problem some other time, and people told me here how to solve it. Any hints?

[devilcoelhodog]

I'm having problems making AutoGK work.

Well, perhaps it is a problem concerning about AutoGK program only, not the subtitles itself.

Now, about subtitles imprinted on the video you can do it using the programs and the tips said before.

Could you have tested one of them? Do you have any feedbback to share with all of us here?

Concerning about OCR issues from subtitles only, AutoGK maybe do not have any relation itself directly. If the subtitles are embedded on the video, you just need to have the file in a proper format / codec that all programs mentioned before can process ( AVI container, for example, for AVISubDetector ) and you can perform the OCR process according to the guides pointed before.

Thanks.

Best regards.


devil (johner)

[carlmart]

I just needed to use AutoGK because one of the programs suggested (AVISubDetector) demanded an avi file, and I didn't have one.

Things didn't go well with Subrip, and I was using an mpeg file with it.

As a matter of fact, AutoGK did convert some of the file to avi, but not completely. As in the past I had used it a lot, and it took about one hour or less to convert a whole voB film to avi, this time I know something is not right.

Anyway, I did try AVISubDetector, but the program is quite dumb when considering dots (on "i") for instance or the dot in "?". and doesn't go on recognizing when you ask it to ignore it, as if it were just a spot. It sticks there and does not go forward.

[devilcoelhodog]

Hello.

You can try one the programs below to convert your DVD content into AVI or something like that:

---> https://www.videohelp.com/tools/sections/dvd-to-avi-divx-xvid

---> https://www.videohelp.com/tools/WinFF ( freeware )

---> https://www.videohelp.com/tools/WinAVI-Video-Converter ( NOT freeware )


If you want to join all your vob / dvd content into one big single vob / dvd file you could test the programs below:


---> https://www.videohelp.com/tools/Mpg2Cut2

---> https://www.videohelp.com/tools/VOBMerge


You could create one big vob file using VOBMerge or Mpg2Cut2. And after that, use WinFF to convert it into AVI, for exemple.

You choose what you think is better for your needs.

My opinion: I would prefer do it all of that manually. Subrip or AVISubDetector never work completely for my needs to OCR embedded subtitles.

It is more "fast" typing and creating all the timing and letter manually instead of testing all the settings about AVISubDetector or Subrip to work Consistently.

Yes, it is more hard, but it's a way to get more consistent results.

You can do all this manually using such program just like this ones below:

---> https://www.videohelp.com/tools/VisualSubSync

---> https://www.videohelp.com/tools/Subtitle-Workshop

---> https://www.videohelp.com/tools/Subtitle-Edit


Thanks.

Best regards.

devil (johner)

[hello_hello]

Sorry, an alternate problem. It has to do with this though.

I'm having problems making AutoGK work. It stops when it gets to Virtual-Dub. It seems to do the processing but then it stalls there and doesn't finish the conversion. I wonder whhat is missing.

I had a similar problem some other time, and people told me here how to solve it. Any hints?

The AutoGK log file might provide some details. It's in the temporary folder AutoGK creates in the same location as the output file. Most of that info is also displayed in AutoGK's main window, but something more specific regardin the exact place the process comes to a halt would probably be helpful.

As a matter of fact, AutoGK did convert some of the file to avi, but not completely.

That seems somewhat unusual. Once again though, details from the log file might help.

[AlvoErrado2]

I use this program: http://sourceforge.net/projects/videosubfinder/

And this guide: http://jumonjigiri.blogspot.com.br/p/extracting-hardsubs.html

No need to convert video to avi, i consider more practical, fast and accurate than other methods.

It is not updated for a long time, but is still efficient.

In a normal DVD, you only need to extract subtitles in format idx/sub, then open this file in Subtitle Edit to turn at some editable format in Subtitle Edit.