VobSub program but for SRT ?

[iNFO-DVD]

Is there another program, like VobSub, that I can run via a 3rd party program, that will make a SRT subtitle file rather than an IDX/SUB ?

[unskinnyboy]

What about SubRip? It takes CLI commands (check CLI.txt in the installation directory) and outputs SRT.

[Matthew]

I can't imagine CLI is of much use as far as srt is concerned. You have to sit there and enter in letters/numbers/characters when prompted. Hence vsrip is much more appealing in my opinion, although the lack of standalone support (by mpeg-4 players) might be an issue.

[unskinnyboy]

Unfortunately, for getting OCR'ed subs as output, I don't think we can escape entering the characters (at least initially), unless we already have a huge character matrix file to start with. Even then it is no guarantee.

[ai4spam]

You can use SubRip's new "Fill matrix from text" feature to save you some typing. Try it out!

[iNFO-DVD]

I had a look at subrip and I must say I wasn't very impressed, though that's probably down to my lack of knowledge on that program. I had a look at the 'CLI' file, tried, but couldn't really get it to work which is down to me, I just got a bit dis-heartened with it all.... and the bloody heat here is killing me ;)

I'm using the command-line of VobSub, which I find excellent, to make an IDX/SUB and refernce it from my AVS file via my program. That's all great as I then burn them directly into the AVI during the encode.

All is/was well.

So I added another approach of having external/seperate subtitle files, so rather than the burnt in subs I had the seperate IDX/SUB files, trouble is, these can be quite massive and the avi size needs amending if you were to say, do a 1 CD convert and the size amendment can be quite significant.

I thought when using external/seperate to either convert the IDX/SUB to an nice and small SRT file or make an SRT file from scratch using the IFO/VOB.

Any ideas on this ? How would you achieve this ? This has to be something I can run without any user intervention.... Is it possible ?

[unskinnyboy]

My 2 bits -

Having to type in the characters (some amount at least) to convert either idx/sub-->srt or ifo/bup/vob-->srt is a necessary evil and can't be avoided since we are converting picture to text (.bmp->text) and no OCR software would be that super-intelligent to be able to read and understand all the subtitle images out there - every font, shape and size. As an example, my current SubRip character matrix is more than 100 DVDs strong and just when I think I got it all covered, there comes a DVD where I have to reenter everything from scratch :\ (Ofcourse, there is the guess character option, but IME sometimes it takes a wrong guess).

My suggestions are -

1) To assume that the user is aware of the size implications when he chooses external idx/sub and hence he will adjust the CD size accordingly (reduce it a bit).

2) To let the user type in the characters for the srt conversion, at least he will get SubRip/SubResync/whatever launched automatically for him if he chooses external idx/sub.

*If* there is a converter which would do all this without user intervention, I would like to know about it too. Because then I can use it as a standalone tool for all my srt conversions and thus end all my troubles. :)

P.S: Would you like to share what difficulties you faced with the SubRip CLI?

Disclaimer: I haven't tested the new "Fill matrix from text" feature which ai4spam talked about.

[ai4spam]

I thought when using external/seperate to either convert the IDX/SUB to an nice and small SRT file or make an SRT file from scratch using the IFO/VOB.

Any ideas on this ? How would you achieve this ? This has to be something I can run without any user intervention.... Is it possible ?
Well, .srt files certainly are nice and small ;).
Unfortunately, no program that I know of does this without user intervention. However, if you use SubRip long enough, you will:
1. get used to it ;) and learn how to use it (I'll write another guide one of these days)
2. end up teaching it a lot of characters, but each of them only once (also, the best guess feature gets better the more characters you input)
3. reuse character matrices: right now I have one that works for 90% of the DVDs with almost no user input, and I can send it to you if you want

[iNFO-DVD]

OK I'll have a little think about it all. Thank you both for your input. (I suppose I was hoping for some magically solution I'd missed).

Thanks again.

[CoNS]

3. reuse character matrices: right now I have one that works for 90% of the DVDs with almost no user input, and I can send it to you if you wantCould you upload it as an add-on to SubRip? Starting with an extensive character matrix would make life a little easier for me :-D

Also, would it be possible to collect character matrixes from different users (for example unskinnyboy) and merge them into one 99.9 % super matrix file?

[ai4spam]

I'll see what I can do about uploading it.
It's pretty clean, I've just looked through it when I implemented extended selections.
As for joining, there is no such functionality right now, but SubRip can be told to look for a character in all the matrices in a directory.

[unskinnyboy]

@CoNS, Here (http://sr2.mytempdir.com/66939) is mine. You will have to unrar it and extract it to C:\Program Files\SubRip\ChMatrix. Hope you will find it useful.

[Matthew]

So I added another approach of having external/seperate subtitle files, so rather than the burnt in subs I had the seperate IDX/SUB files, trouble is, these can be quite massive and the avi size needs amending if you were to say, do a 1 CD convert and the size amendment can be quite significant.

You can rar the sub file with max compression (although there could be some licence issues there). DirectVobSub will read directly from the rar during playback. That gets the size down by a lot.

Anyway if you are going to proceed with srt support then you'll need to code your app so that users have to respond to subrip's prompts.

[ai4spam]

Another thing you can do is strip the languages you don't need from a .idx/.sub file pair, and then RAR them. I forgot which program does that, maybe vStrip?

Anyway, here's my char matrix: http://sr2.mytempdir.com/67113 (thanks unskinnyboy for the link above, I didn't know where to post it).
It has 8000 characters.

Additional details here:
http://forum.doom9.org/showthread.php?p=679590#post679590

[unskinnyboy]

Another thing you can do is strip the languages you don't need from a .idx/.sub file pair, and then RAR them. I forgot which program does that, maybe vStrip?
VobSubStrip (http://www.free-codecs.com/download_soft.php?d=787&s=403) + SubResync will do that. Here (http://www.doom9.org/vobsub-editing.htm) is the Doom9 guide. However this requires considerable user intervention, like editing the .idx etc.

[ai4spam]

Doesn't look too complicated to me, it says 1a is an alternative for 1, so you can just use VobSubStrip instead of editing yourself.

[CoNS]

ai4spam and unskinnyboy: Thx to both of you for your matrix files :-D

[ai4spam]

One is glad to be of service ;).
<and some dummy message here to make the forum accept this post>