JDPFu.com 2025

I’ve been working a project with a friend – he shall remain Nameless at his request. Anyway, we were playing with some changes to VLC software related to subtitles and I mentioned the issue getting CC3 for a telenovele that I’ve been watching. We were in the VLC code where subtitles are handled, so I had hope for my CC3 dreams. I could add a few printf() calls and have the CC3 data dumped from VLC. Of course, VLC handles all sorts of subtitles and closed captioning standards pretty well, but using it is not always an option, since it doesn’t work on all platforms. Hacking VLC for this purpose is a major hack and would certainly be rejected by the VLC team. I know that I’d reject it.

A few hours later the same evening, I get an email from Nameless with a English SRT attached, but no explanation. He likes to do that. It is after my bedtime, so I figure the explanation can wait until later. I’ve been trying to solve this for awhile.

Explanation

The folks that do Closed Captions are TV/video people. They are do not seem to be computer people. Someone who bridges that gap is pretty rare – lucky for me, Nameless is just that sort of person. He recognized the CEA-608 NTSC names and the use of Field 1 and 2 inside some code. He became interested and did some quick googling and found a program that handled subtitles and captions. That program had an option to dump field 2 text in line 21 streams. He used it with a few options and ended up with CC3 dumped into an SRT file. Perfect.

The program? ccextractor – yes, the program that I’ve been using for years. I’ve pulled the code and searched it for CC3/cc3 and never found anything. The support website makes reference to CC3 patches that were lost. Those two datum were enough to convince me that it wouldn’t have the desired capabilities. Obviously, I was wrong.

Hopefully, this outline will help explain the CEA-608 organization:

CEA-608 NTSC

Line 21

  Field 1

     CC1

     CC2

     T1

     T2

  Field 2

     CC3

     CC4

     T3

     T4

     XDS

I Feel Stupid

Nameless didn’t have the same preconceived ideas. He simple pulled the ccextractor code down, compiled it, and read through the help. He was able to recognize that an available option already existed in the code to pull data from both Field 1 or Field 2, separately or together. CC1, CC2 are field 1 and CC3, CC4 are field 2. A quick test and he knew it worked. All that remained was sharing his success with me.

I had pulled the latest source for ccextractor myself and compiled it using the unusual build script. I had already decided to help the project out by creating a makefile, cleaning up the multitude of compiler warnings and building libraries for the parts where that would make sense. The code appears to be a hodge-podge of C and C++ written by many different people over the years. It works and the feature set is pretty awesome, but the code is not what I’d call pretty. I still need to contact the core dev team to see if they are interested. If the changes will be rejected, this isn’t useful. I’ve helped a few companies migrate their source code to more maintainable setups.

ccextractor is cross-platform, so any changes risk breaking the code on MS-Windows, OSX, Linux, BSD, Solaris, and even ARM devices.

The CC3 Dump SRT Command

To extract CC3 from an MPEG2 recording, use

$ ccextractor -2 -cc1 tv_show_file.mpg

To extract CC1 and CC3 from an MPEG2 recording in a single command, use

$ ccextractor -12 -cc1 tv_show_file.mpg

To get CC4 data, I think this command should work

$ ccextractor -2 -cc2 tv_show_file.mpg

Other options that I use, are;

—nofontcolor -utf8

Adding SRT to MKV Files

Getting the SRT files out isn’t my end goal. I love MKV video containers as you can see. Putting multiple captions, subtitles, and other files into the MKV container with the audio and video streams is my end goal.

I use mkvmerge to do this in a script. The basic format is:

$ mkvmerge -o output.mkv input.h264 english.srt

Add multiple SRT files to a single file.

$ mkvmerge -o output.mkv input.h264 spanish.srt english.srt

$ mkvmerge -o output.mkv input.h264 —language 0:esp spanish.srt —language 0:eng english.srt

Perfect. I’ve used descriptive file names, not the actual filenames output from ccextractor. Of course, ccextractor does let us specify the output SRT filenames, if you want to control them.

So, I’ve finally gotten CC1 and CC3 out and will get them added to MKV files with the correct language labels.

I still need to suggest documentation changes to the ccextractor team. Hopefully, they will expand the explanation into the README docs and the help screen for the program.