Capturing Closed Captions from an AVI

[merlin001]

What I'd love to do is capture from my Dish to an AVI, pull the Closed Captions off of it, convert them to subtitles, and be able to use them in a DVD. So, any ideas here? If I can't, I'll be doing them all manually, which, quite frankly, sucks.

Thanks!
David

[dphirschler]

I don't know of any automatic tools that can do this, but I have read of a filter that can do OCR on the frames (if you actually captured the video with CC visible), and get you a good head start. I can't think of the name of the filter though. Look on Donald Gaft's website. It should be there. The downside is that you will have to capture a second time with CC turned off.

I would recommend capturing directly from satellite into AVI with CC turned off while also recording to VHS. Then playback the VHS for the CC capture. It will still be a pain, but may be the best way so far..


Darryl

[dphirschler]

Here's an update.

Sublog - Virtualdub filter which extracts hardcoded subtitles from a video stream and creates an index file

Subrip - Stand-alone application which processes the images (using OCR) into text subtitles which can be edited and manipulated.

Links:
http://www.xs4all.nl/~vielle/video/sublog.html

http://www.subrip.fr.st/

Good luck. Don't forget to write about your experiences. I plan to dive into this someday, but not until I finish my current project.


Darryl

[handyguy]

Being deaf I do this alot however my way is easy. I just start a dvd with subtitles (Amost 90 out of 100 use them instead of captions now but call them captions on the box anyway), input to my dvd recorder, which creates dvds & presto, open captions on the resulting dvd.

all in wonder video cards can capture captions if you want to from any video source & write to html, word, etc

[deadpac]

merlin001,
I don't think it is possible to extract the line 21 field 1 data stream (Closed Caption 1&2 and Text 1&2 areas of a broadcast signal) from an avi file. You can however capture the Closed Captions stream from a TV show as it's being aired or from tape (DVRs such as Tivo, ReplayTV and DVD recorders probably do not capture the closed caption stream as they are probably just capturing and compressing on the fly and probably would be out of the question; but dont' quote me on that, does anybody with any of these devices know if Closed Captions work?). This requires capturing the show first to get the video and audio; and then going back again to capture the show a second time for the Closed Caption Stream. Capturing the video is totally up to you, it doesn't even matter if you capture to avi or mpeg; although starting from avi will make it easier to sync the closed captions to the video.

The only way I've seen to capture Closed Captions stream from a video signal; being broadcast analogue, cable, sat, dig-cable.......etc is by using Graphedit. Graphedit is a very unwieldy program that can be downloaded from microsoft.com and is part of the tools that come with directx. I just recently got some info from a very helpful member of these boards. Getting your TV card to do what you want in Graphedit is an excercise in patience and sometimes futility. Even if you do get graphedit to do what you want and you've captured the closed captions stream, it is a raw file and getting them from a RAW file to something usable is another bout of frustration waiting to happen because usually the header file is either corrupt or missing.

You would think capturing something like closed captions would be easy, but quite the contrary. A lot of people like handyguy use the ol workaround of capturing the captions with the video. While that may work is does have 2 really big drawbacks; one being it's permanent and can't be turned off and secondly if you don't encode you're video properly you might end up making the captions unreadable. There are almost no guides on the net to do this and even if you find something to do with closed captions it's probably about ripping them from DVDs; well that doesn't do a lot for the everyday television programs people watch.

dphirschler,
I've played with sublog with various sources and it looks like it looks for visual text on the screen and attempts to interpret the letters as they are seen on screen; this probably isn't going to work to well seeing as how a lot of capture programs don't even allow you to display and capture subtitles as part of the video. EDITED just re-read your post and you mention this, sorry.

merlin001,
If you really want to try your hand at closed captions, it is possible. It involves using Graphedit and you would end up having to capture the show a second time although the second time around you're only going to be capturing the raw CC stream which FYI for a 2hour show is only about 300kb; so don't worry about file space. I would recommend taping the show so you can get the cc stream. It also wouldn't hurt to know a little bit of hex editing just in case you get everything to work but can't convert the raw CC file because it's header is either missing or corrupt. Then taking the time to sync the subtitles with the video which is very tedious and sometimes frustrating because of the commercials with captions. After syncing comes the correcting of errors; a good 90% of broadcast closed captions are in ALL CAPs and special tags that can't be tranlated plus italics and punctuation errors.....etc Then converting to DVD subtitles and using a DVD authoring app that supports subtitles ; which by the way is far and few between most notably IFOEdit(basic), Adobe Encore (midlevel), DVD Maestro and Sonic Scenarist (commercial level authoring).

some notes:
-Graphedit can manipulate anything that calls direct show routines and can be very dangerous to the equilibrium that a lot of users enjoy when they get their computer to to do what they want and have it tarnished as soon as that decide to play with something new; for instance playing around with Graphedit and not knowing what I did caused my Windows Media Player not to be able to decode Mpeg2 video anymore and the applications that came with my card where unable to detect the card anymore; luckily a bootable Ghost CD and a recent image are never far from any of my boxes; and it shouldn't be for anyone else who likes to play around with their computers.

-I to this point after only a couple days of experimentation have only been able to author one dvd with a tv show to dvd with selectable subtitles that came from captured closed captions. Using just IFOEdit means no flexiblity and you're not going to be able to use a menu system. I'm trying to figure out Adobe Encore but the type of subs that it supports are very hard to find information on how to make bmps for those formats; and plus I really try to avoid Adobe products at any cost. Having finished college I do not want to get to the point where it feels like I'm learning another programming language in order to figure out how to use DVD maestro or Sonic Scenarist; I've read on various dvd authoring forums of people who have read entire dvd authoring books have not been able to figure out how to use Scenarist to its fullest potential.

-it's easier to go to svcd or cvd with selectable subs from closed captions than it is to go to dvd with selectable subtitles form closed captions. I have made several svcds (and cvds) with selectable subs from closed captions using tv shows and some vhs tapes as the source.

-I unfortunately lost an 80gb to causes unkown; and it was the only drive I had that contained the detailed information about using Graphedit to capture Line 21 field 1 data aka Closed Captions. I still have the screenshots of my working Graphedit setups in windows 2000 and windows xp. There is a slight difference betweent the two when it comes to using Graphedit to capture CC streams in those two OS's. The webpage that the member found this information on is unbookmarkable and very hard to find, but it is out there.

Still interested?

I'd be if you were still interested based on the way I express enthusiasm in a subject; usually with a dark outlook. Also not to mention the amount of work it takes to achieve very little (depending on your view of subs).

Wish I could help you but even my screenshots and the working GRF files I had are also mia thanks to that dead drive. I give really horrible explanations and I probably wouldn't even recommend getting info from myself.

[uspino]

Great explanation. Only one thing may be added:
Apparently, Hauppauge is refining the drivers for their PVR250 (the most widely available hardware capture card) for CC to DVD functionality. There's already a registry line named InsertCConDVD=1 that is suposed to convert CC text to DVD subtitles on the fly and add them as such into the DVD stream. Seems too good to be true, so let's wait and see.
In fact, Windows Media Center, using the same card, is already filing the CC text into the recordings.

Please, keep this thread going.

[spiderman2k1]

This might help https://www.videohelp.com/forum/archive/t135258.html you can get a ATI AIW from www.pricewatch.com for $33.00 dollars get a Radeon. Might want to get a DVD recorder they record Close caption just like VCR do. No extra work to do like on the PC.

[deadpac]

wow

I didn't even know I stumbled upon the page that kalayaan was referring too when she said that she could not bookmark the page for the information on how to capture closed captions from an analogue source.

http://www.sonic.net/~mcpoodle/SCC_TOOLS/DOCS/SCC_TOOLS.HTML

That is the page with the detailed instructions on how to capture closed captions using Graphedit. I was completely oblivious to the fact that the instructions that Kalayaan sent me where right ther in the middle of that page. I used that page to only download the SCC_Tools package because Kalayaan had already pm'd me the instructions for cc capture with graphedit. All I had to do was find the raw2scc.exe tool she was referring too and using just plain ol google is how I stumbled onto that page. *Hits forehead*

spiderman2k1,
I forgot what keywords I used to find that post but that is the same post that I used to get into contact with Kalayaan, the very "helpful member of these boards" that I was referring to in my post above.

ATI AIW are nice but they don't really do much but give you a transcript of the broadcast in Word format and with no timecodes, can't really do much without the timecodes. I didn't know DVD recorders could record closed captions; got a friend with a set top recorder, I think it's time to pay him a visit and check that out for myself.

uspino,

Seems too good to be true, so let's wait and see.

Only a few authoring apps support subtitles. Even then there are some tradeoffs when it comes to using these applications, hopefully soon those popular dvd authoring apps such as Ulead DVD Workshop, TMPEG DVD author, DVD-Lab,....... will soon begin to support subtitles.

one new years resolution for me: Stop shying away from Adobe DVD Encore and figure out what the hell it wants