[FFmpeg-devel] [PATCH 2/2] lavd/lavfi: allow to extract subcc.

anshul anshul.ffmpeg at gmail.com
Tue Mar 18 08:04:56 CET 2014

On 03/17/2014 12:24 AM, Nicolas George wrote:
> Le quartidi 24 ventôse, an CCXXII, Michael Niedermayer a écrit :
> First, lavd/lavfi is maintained by Stefano; he seems busy these days.
> Second, there is the issue of the name of the codec and its id. I called it
> CEA-708, but Anshul says it should be CEA-608; unfortunately, I do not find
> any trace of such a thing, there is only EIA-608. According to wikipedia and
> Reimar, CEA-708 would be a combination of EIA-608 and some more data; but
> maybe that is already what is coded by AV_CODEC_ID_EIA_608.
Sorry, I discussed this issue on irc but forgot to mention,
Peoples have already answered that CEA-708 can contain the CEA-608
data. As your patch discover the data as 708 and that does mean that
it may contain 608. So I think this patch is doing good about CEA-608
as 708 since you have decided the data is 708 after analyzing starting 
even those frame were 608  and after sometime if get 708 frames
then it might not break anything.

> Third, there is the issue of testing. Right now, this patch allows to just
> dump binary data, and an ASCII/hex dump of that shows promising tidbits of
> text, but I would be much more comfortable if someone would produce code
> that would allow to do something actually useful with it.
I am working on it.would update soon.
> For the last point, I believe Anshul can be of some help: you gave a small
> source code that would save the same binary data in a file. Can you explain
> what you do exactly with that file afterwards? I do not suppose you can just
> give it as is to ccextractor.
CCExtractor support raw data of closed caption, and the data that is dumped
by our code is raw data of closed caption.


More information about the ffmpeg-devel mailing list