Super User is a question and answer site for computer enthusiasts and power users. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

When I open a .sub file with SubResync, it shows the subtitles of a video file in "images":

enter image description here

enter image description here

Basically, I want to convert those subtitles into text form. However character recognition converters usually only allow image files (e.g. pbm, bmp, jpg).

Is there some way to convert a .sub file into an image file, or a couple of image files?

share|improve this question
up vote 2 down vote accepted

You don't need to convert your .sub files into images, SubResync has built-in OCR capabilities to convert them directly to plain text.

From this tutorial:

Click open, goto to directory that has your filename.sub

enter image description here

Select your filename.sub by single click then open. By default fps is 25, if your video frames per seconds a different value (23.97, 24, 25. 29.97) select something else.

enter image description here

The subresync dialog box default "Render" checked and "unlink" not checked. From the "drop down menu" it may detect multiple language options(Eq. 00-Japanese, 01-English, 02-Chinese, 03-Not detected, 04-etc.....)
I only comprehend english so 01 was selected.

enter image description here

Then clicked Save As, from the "save as type" drop down menu, select your desired output extension.

Note: OCR Setting "clear image->letter(s) database" will build a database of letters you will be typing. If checked this item will clear the database. Click save to begin.

enter image description here

Follow the moving red box below and "enter matching letter/word/(s)." Once letter is typed press enter. This in turn builds a database of the letter/words/(s) you type. Some times the words don't appear properly cause of unknow character or symbol. Click "extend" to find the remain letter/word/(s).

enter image description here

Finally test your new and see if matches with your flix.

share|improve this answer
This is a powerful answer thanks =) But it doesn't work as well for languages like Japanese and Chinese (yes it's better than typing them out individually, but there's like thousands of different letters/characters..) Is there anyway to simply convert it into into an image file so it could be fed into a character recognition program? – Pacerier May 28 '12 at 21:55
@Pacerier Short of taking a screenshot of each entry in the .sub file, I'm not aware of any way to do that, sorry. – Indrek May 28 '12 at 22:26
@Pacerier you are feeding it into a character recognition program. In this case, you are simply teaching it in the beginning what these characters stand for. If you were to do this with English subs, for example, you could come across numerous examples where you would be presented with two letters that the program recognized as a single character (because of proximity, pixelation, etc). As you progress with what you teach the program, it automatically converts more on it's own. – Bon Gart May 29 '12 at 0:14
@BonGart Yes that's the problem, English has only 26 characters so it would only be a matter of minutes to "teach the program". Japanese has more than a thousand characters so it would take hours to "teach the program". If I could convert it into an image, I could feed it to an (already-self-taught) dedicated Japanese character recognition program which would output the Japanese characters. – Pacerier May 29 '12 at 1:25
@Pacerier I see the dilemma, but you would still run into the issues of what the character recognition program with a pre-programmed database would do with the characters it could not recognize. – Bon Gart May 29 '12 at 1:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .