Closed Captions to .srt, too many errors!!!

[joseph5]

I have a standalone DVD recorder(LVW-5005). It can record and display closed captions from the TV fine.

I want to convert them to .srt file format. First, I tried VSRip, but it gave an error. Next, I followed this guide:
http://www.geocities.com/mcpoodle43/SCC_TOOLS/DOCS/DVD_FILTER.HTML

I got the .bin file and converted it using CCParser to .srt, however, the subtitle file has too many errors:
1
00:01:07,500 --> 00:01:09,299
A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@@@g@(

2
00:01:09,299 --> 00:01:09,900
A@AR@@E gYOU SU@RE( YOU'VA@E CORRtDEC{}TEveDr

3
00:01:09,900 --> 00:01:10,466
JK
P!vTHE @@PRgOBLEM,@ D(R. SA@IN@@CLgAIR? @


4
00:01:10,466 --> 00:01:13,599
(A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@@@g@(A@tD{}ve}ezKPU@@g@(A@@@g@(A@

5
00:01:13,599 --> 00:01:13,933
@@
g !AB@SO(LUTELYA@,@@


6
00:01:13,933 --> 00:01:14,266
g
!MR@. (LUTHORA@.@@ g

The original captions don't have all those strange characters when playing the DVD on the recorder.

Anyone knows what I did wrong?

Thanks in advance.

[CoNS]

You may find some inspiration here: http://forum.doom9.org/showthread.php?t=107148

[slk001]

I haven't had much luck with CCParser giving a good (accurate) CC stream, but received results much like yours (albeit, maybe with less errors!).

I just converted a DVD's CC stream to a subtitle by using McPoodle's SCCTOOLS to give me a text stream with timing, then importing the results into EXCEL to get rid of the CC control stuff, then adding what was needed to bring the data into SubStationAlpha (what I use), then just retimed the whole thing. Reads more complicated than it actually was.

[joseph5]

You may find some inspiration here: http://forum.doom9.org/showthread.php?t=107148

I followed the instructions given there with "raw2scc" and "ccasdi", but now the file is worse. Even the times are incorrect:

1
00:00:00,000 --> 00:01:56,801
|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ|ñ

2
00:01:59,167 --> 00:02:06,701
JK
PTH

3
00:02:30,634 --> 00:02:37,267
@@
AND @
WAS

4
00:02:37,334 --> 00:02:41,801
@@
MR

5
00:02:41,868 --> 00:02:46,467
HK
PI ME@@
OF

6
00:02:53,868 --> 00:02:57,801
YO

7
00:02:57,868 --> 00:03:02,467
@@

Any other ideas?

[McPoodle]

If you could post the .bin file somewhere, perhaps I could figure it out.

[joseph5]

http://rapidshare.de/files/29533505/capt01.bin.html

Thanks!

[McPoodle]

Whichever engineer was responsible for closed caption encoding for your LVW-5005 DVD Recorder should be taken out and shot. Repeatedly.

It didn't help that the captions were also in the wrong format.

I tried to come up with some tool that would allow you to fix these files on a regular basis, but I'm afraid that isn't possible. Buy a Panasonic if you really need to convert captions to subtitles in the future.

You can find what I managed to recover at http://www.geocities.com/McPoodle43/SCC_TOOLS/capt.srt. I would not be surprised in the least if the timing was way off or if some subtitles were missing, but at least now it's readable.

Sorry I couldn't be of more help.

--McPoodle

P.S. Here is what I did to get this, if you're willing to go through that much trouble for each file:


Download rawproc.exe (http://www.geocities.com/McPoodle43/SCC_TOOLS/rawproc.exe) (just updated) and split.exe (http://www.geocities.com/McPoodle43/SCC_TOOLS/split.exe) from my site and put them in the directory that has capt01.bin in it. You should also have raw2scc.exe, scc2raw.exe and ccasdi.exe from SCC_TOOLS in the same directory.
If your .bin caption file is not called "capt01.bin", then rename it to "capt01.bin".
Copy the following into a text editor and save it as "convert.bat" in the same directory as everything else (lines 2 and 3 appear to be redundant, but I needed them to get this to work):
rawproc capt01.bin capt02.bin
raw2scc capt02.bin
scc2raw capt02.scc
split -h4 -s2 capt02.bin
raw2scc capt02_odd.bin
raw2scc capt02_even.bin
ccasdi -s capt02_odd.scc capt02_odd.srt
ccasdi -s capt02_even.scc capt02_even.srt

Double-click convert.bat. When it's done, you will have a few more files in the directory, including capt02_odd.srt and capt02_even.srt.
Open both SRT files in a text editor and copy and paste the lines that make sense from each into a new text file. The pattern I saw was that you'd start with subtitles that made sense in the _even file and a little bit of garbage in the _odd file, then there'd be a commercial break, and when the show came back on, the subtitles would be in the _odd file. There was no swap at the second commercial break (just a few garbage lines in the _odd file I pulled out), and once there was a swap in mid-sentence for no reason I could determine.
Save the new file as capt.srt and delete the other new files.

[joseph5]

It didn't help that the captions were also in the wrong format.

What do you mean? When I ran your script, rawproc gave the following error:
Input in DVD format--this program isn't necessary.
Do I need another tool to convert it to a format that rawproc will work with?
Sorry I couldn't be of more help.
Actually, I'm very grateful. Thanks a lot!!!

[McPoodle]

You did download the new version of RAWPROC from the link in my post, right? When you run it by itself, does it give a version of 1.1? If you're still having problems, then maybe there's a caching problem with my site and I'll try putting the file somewhere else.

As for the format issue, I could fill pages with what's wrong with that file you sent me. A lot of a raw DVD caption file is made up of filler bytes. All DVDs and most DVD recorders use a filler byte of 00, which is never used for anything else, so it's easy for me to skip it. The LVW-5005 uses FF, which is the same byte used to flag where the actual caption begins. Next, DVD captions are on two channels, one for the text, and one for the announcements ("You're watching channel 13." "You're watching Married with Children, which started ten minutes ago and will end twenty minutes from now." "The time is 7:16 and 20 seconds." And so on...). The announcements repeat several times every second and make up most of a closed caption file that isn't filler. In a normal file, the captions and the announcements get different flag bytes before them, so it's easy for me to strip out just the captions, but the LVW-5005 puts the same flag byte (FF again) in front of both, which is why you get so much garbage in the files even using my method.

[joseph5]

Thanks for the explanation.

The file I have is version 1.0. I downloaded it again from the link you posted above, but it's still 1.0.