Saving Broadcast TV Subtitles

[sammie]

I was wondering if anyone new how to rip and save subtitles from a broadcast tv program? I know they are still there on a tape or TiVo recording, but the seem to disappear somewhere in processing to an MPEG file. I haven't figured out how to see if they are still there in the avi capture yet, but was hoping someone else has tried to do this. Yes, I know I have to use the TV receiver CC function and not the DVD player.

[snowmoon]

I would like to figure this out as well. I believe it's stored somewhere in the refresh area of the signal and therefore is not transfered onto the mpeg since that only hold the visible are of the picture ( NTSC is 525 lines while mpeg only contains 480 ).

And just to clarify some things...

Closed Captioning is a signal that is integrated into the video and is decoded by the TV. They are stored as charater data.

Subtitles are specific to DVD's and are stored as picture data in the mpeg stream.

[sammie]

Right about the differentiation Snowmoon.

I did find this link at Doom9
CCParser 0.05b - parses Close Captioning DVDs http://www.doom9.org/Soft21/Subtitles/CCParser0.05b.zip

This is supposed to strip CC stream and save as some sort of DVD subtitle information. Seems like it should work and implies the CC info. should be there in mpeg files. I know it doesn't always seem to be in mine. I've actually only tried 2 or 3 so I can't say for sure what %. I do often trim off the noise bands before encoding so based on what you said I wonder if that might be the cause? Unfortunately, I think the program I listed above only works on mpeg, otherwise it would be easy to capture, separate CC track and audio, then encode and and add the audio and new subtitles back in. Hope I can find some answer. My wife has problems hearing dialog sometimes, and likes to use subtitles on dialog intense movies.

[sammie]

....missed a detail in my post. The program author wrote it to get CC info from a DVD that used CC and not subtitles. Therefore it seems like the info must be ina part of the video stream that is maintained in mpeg format. Found a bit more information that the program looks for the information in "Line 21". Maybe that will help someone who has more video signal knowledge than me to give an answer.

[snowmoon]

Since mpeg ntsc is 480 lines and NTSC is 525 their are 22 or 23 lines lost top and bottom of the signal. So line 21 is never captured nor encoded. At least that's the theory.

I would love to know if it can be done. It would be a great addition to my TV cap collection.

Cheers

[donpedro]

What would happen if you capture by pass-trough device (DV-cam) from TV with CC turned on ? Just a thought.

[sammie]

@Snowmoon. Jives with the "line 21" bit. Have to wonder about how it was in a DVD stream though? Would argue that even if it is captured somehow, my habit of cropping would not be a good thing.

@DonPedro. The question is how do you "Turn them on". That's a funcion of the TV. Plus, I don't wan them on all the time. Want to be able to turn them on an off like traditional CC or subs. You do raise an interesting question though. If you captured to DV, then imported the DV via firewire, is the info still there? I've never looked.

[snowmoon]

sammie:

My understanding is that CC data is stored as a file on (S)VCD's and added to the video stream by the player. I'm not sure if it's that was for DVD's or if it's another type of stream that's multiplexed with the file.

CAPTnn.DAT file.
For each track containing user data with Closed Caption information a file shall be present Other (CD-ROM) files may be included in this directory, usually intended for computer based systems.

from http://www.cdrinfo.com/Sections/Articles/Specific.asp?ArticleHeadline=SuperVCD+Format&...ries=0&index=1

Cheers.

[lewijo]

Here's a concept for further discussion:

My ATI All-in-Wonder can save the CC data in a file. The file could (conceptually) be reintegrated into the video thru the subtitle facility of a DVD authoring program.

This might require that all of the text be converted into bitmaps. I don't know, I'm just starting with authoring and haven't looked at it's capabilities. And who knows what kind of manual effort this would take!

Just an idea.

lewijo

[donpedro]

Well... my method would be permanent. (Just theory) If you capture yout OUT signal from TV video (If CC is turned on TV) would have CC in it. So you will end-up with video that have CC permanently as part of it. No hidden in signal.

[donpedro]

I believe that DVDs have (some) also CC on it not just Subtitile. I know this because my wife doesn't like Subtitles so we watch movies from DVD with CC-On. I am not sure if it have to be as function of DVD-player, but my Mintek-2580 does that. (Can't be connected through S-Video, must use standard video output)

[sammie]

@lewijo. Were you using the AIW tuner or the line in part of the capture. Could see how that would make a difference.

@DonPedro. Yes, per my web searches it does sound like some DVDs come with CC not subtitles. This would seem to imply that CC is not totaly incompatible with mpeg (or at least DVD mpeg). It is just for the few I've looked at, it doesn't seem that what I'm doing (VDub capture and processing TMPEGenc encoding) leaves the info intact.

Since it sounds like all of us are guessing right now and that this seems to be very much one of those "Dunnow, never tried things." I'll do a few experiements using the program I found and the DV stuff DonPedro suggested and see what I can come up with.

[Cornucopia]

This is what I understand about CC info.

*It only exists on NTSC (PAL does it much differently).
*Ususally resides on Line 21 of the Vertical Interval (non-displaying on normal monitors)
*Uses a pulse-type code that normally gets decoded only at the TV.
*Exists only when all 525 NTSC lines are allowed to pass thru unmolested.

Therefore, if you cap 601/D1 (486 lines) or DV (480 lines) or DL something that has been capped that way (or less # lines) then you will have already LOST the CC info.

How does it get onto DVD then? Well, there are certain Hardware MPEG encoders (I think the Spruce Maestro MPX3000 does?) that separately read Line21 and encode CC info into the MPEG stream.
Where does it get encoded? It is AFAIK put in the UserData section of the Muxed System/Program Stream. That's why CCParser can find it and convert it.

DVD players that support CC info (not all do--I found my Apex AD-600A doesn't) will read the UserData section and regenerate Line21 on the fly for composite output.

BTW, CC stands for Closed-Caption. Intrinsically that means it's switchable on/off by the user. If it's been burned in to the picture it is now called Open-Caption. Best place to learn CC info is prob Boston's WGBH PBS-TV station website.

HTH,

Scott

[donpedro]

I'll do a few experiements using the program I found and the DV stuff DonPedro suggested and see what I can come up with.

Did you have chance to experiment ... ?

[sammie]

I did with mixed results. It gets lost some (most times). I haven't detected a pattern yet. Of course part of the problem is that you have to go through the entire process, all the way to authoring and burning) to test it. Hard to come up with ways to separately test capture, edit, encode, authoring, etc. Using Studio 7 to dump back to tape seems ok, but what if that is part of the problem????

[spikeylikesit]

When I did my first DVD rip using SmartRipper and then burned the movie to DVD the captions were there, so it must be in the files (all subs were there too). As far as the capture from broadcast- I concur that it must be in the lines not captured.