Manually ripping subtitles from TV shows

[encodedprocess83]

I have several anime DVDs that I ripped to .mp4s by episode via handbrake. They have both english and japanese language tracks, and i would like to rip the subtitles off the dvds (i still have the VIDEO_TS folders ripped to my hdd) and have soft subtitles for whenever i decide to watch in japanese with literally translated subs. I'm not quite satisfied with any of the OCR applications, as there are so many errors scattered throughout, that I have to go through the whole thing any way. To add to this, the episodes are separated by chapters, not titles. what im looking for is the following:
1) an application that allows me to manually type the subtitle text (while displaying the subtitles encoded on the DVD) that automatically marks the in/out timestamps of the subtitles.
2) an application that will split the resulting .srt file into pieces that correspond to each episode, based on given start/end points, etc.

I would like the resulting file to be a .srt because I plan on using Muxo to merge the subtitles onto the mp4s. (if theres a better alternative, please feel free to tell me, but the priority is ripping the subs)

I run both mac and windows, so any combination of apps is fine by me.

[manono]

On the one hand you say:

I'm not quite satisfied with any of the OCR applications, as there are so many errors scattered throughout, that I have to go through the whole thing any way.

Yet on the other you say:

I would like the resulting file to be a .srt because I plan on using Muxo to merge the subtitles onto the mp4s.

So I don't understand. Any mistakes from an OCR (assuming manual typing, and not those crummy automatic OCRs) is due to your own shortcomings as a typist, and not the fault of the software used. In any event, it's always a good idea to at least spell check and do some minor editing afterwards. Have you tried SubRip or SubResynch. To avoid the typing and to keep the DVD subs, why can't you use VobSubs?

[encodedprocess83]

What I meant by not being satisfied with OCRs is that applications like SubRip often result in misspellings and random characters floating around. I haven't done any manual typing, aside from the recognize-this-character-so-subrip-can-continue boxes, which I'm sure I didn't screw up. My original intention was to completely skip the automation altogether and, line-by-line, retype the subtitles manually. The only automation I want is for the software to keep track of the timing of the subtitles.

Basically, since there's an image associated with each display of subs, I want an app that displays these images in order, marks a .srt when they enter and leave, then prompts the user to type exactly what is displayed and directly place that into the .srt-- no automatic character recognition whatsoever, so that if there are any errors, they would all be MY fault, as you say.

I've tried SubRip, not too familiar with SubResync... after rereading my intention, would that do it? I briefly looked through its description and it seems it does basically the same thing as SubRip... I'll look into it.

As far as VobSubs go, I wasn't aware that they could be used for on/off subtitles-- just hard encoding... I may be mistaken. Will they work with on/off support in a .mp4 container, and play in most software? (the ultimate destination is a media center powered by Boxee, but also possibly iPod touch/iPhone) If that's the case, how would I go about merging them with my mp4's?

[manono]

As far as VobSubs go, I wasn't aware that they could be used for on/off subtitles-- just hard encoding... I may be mistaken. Will they work with on/off support in a .mp4 container, and play in most software? (the ultimate destination is a media center powered by Boxee, but also possibly iPod touch/iPhone) If that's the case, how would I go about merging them with my mp4's?

They can certainly be used as selectable subs. Many use them with AVIs. I don't work with MP4 and have no idea even what Boxee is, so the only way to be sure (unless someone else with more experience with those than I shows up) would be to test yourself. Ordinarily, to make the VobSubs play with the movie you name them all the same - Movie.mp4, Movie.idx, Movie.sub. When I'm back at my encoding computer I'll check it out. I don't have any idea about muxing them into the MP4, but will test with the subs kept external.

Although I have no idea why you're having trouble with SubRip, I think if you play around with the colors (choose different ones than what SubRip offers), you can get a box with the dialog and then you can type in every single line.

Edit later: I just tested changing the colors. You push 'Start' and then 'Change Text Color'. Fool around with it for a minute until you get a combination that forces you to type in the entire lines. It'll say 'THIS SUBTITLE ITEM CAN'T BE PROCESSED' and have a space for you to type in the line. Then you'll get the original timings together with whatever you type in. Still a very bad idea and a waste of time, if you ask me.

I also tested playing external VobSubs with an MP4 video and it also works fine. I don't know about it working after being muxed into the video, or even how to mux VobSubs into the video. I'm sure someone around here knows those things, though.

Edit Again: After looking around, I'm fairly certain the MP4Box and YAMB GUI combination can mux VobSubs into MP4:

- subtitles with srt (SubRip), ttxt (GPAC), sub (MicroDVD), idx/sub (VobSub) and xml (QuickTime TeXML) as input formats

http://yamb.unite-video.com/features.html

[DereX888]

SubRip is excellent tool, but it doesnt work well "out of the box".
Actually you have to "fine tune it" to each DVD's subtitles, either with colors or by teaching OCR new characters.
It is very good OCR scanner, but it do what you "taught" it only.

Depending on fonts used in the subtitles, usually problems arise from similar letters like "I" and "l".
Other than that, if you have too many character recognition mistakes (like "random characters" you've mentioned) it probably means you were using character set from another DVD's subtitles. Just "teach it" the characters instead of using previous .sum file, that's all.

[encodedprocess83]

So I played around with everything suggested and ended up going the way of external vobsubs. Using YAMB, I muxed a test episode with sub/idx files, and this played wonderfully on VLC, but Boxee, QuickTime (and, therefore, iPod) failed to recognize/play the additional stream [I'm not surprised about this-- QuickTime always had horrible compatibility with files]. When using them externally Boxee was happy to play them when I wanted them to.

I also tested converting to .srt manually, and automatically. Both of you were right-- it is indeed a pain to go through it line by line... I don't quite know why I wanted to take on that job, and adjusting the colors yielded better accuracy as well. When muxing these .srt's with Muxo, they played as expected on everything I wanted. I ultimately decided to go with external vobsubs, however, because it's not especially necessary to get these done immediately, and external vobsubs already solve the problem with little human intervention-- plus, I can choose to convert them in the future if I'm *absolutely* sure it's desired. I probably won't end up watching these in Japanese on the go anyway.

Also, if anyone stumbles upon this thread with a similar question, the second part of my question (splitting the subtitle files), if you go the way of vobsubs, VobSub has a nice tool called VobSub Cutter that can do this quite well. It will require you to know the exact time that chapters begin and end, but that's easily found using any media player.

Thanks guys!