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

Nicolas George george at nsup.org
Wed Mar 26 17:19:37 CET 2014

Le septidi 27 ventôse, an CCXXII, Carlos Fernandez a écrit :
> CEA-608 and CEA-708 and the new names of the EIA-608 and EIA-708
> documents. For 608 the exact title in the first page is
> "ANSI/CEA-608-E" (for the revision I have, anyway).
> 608 describes the 'old' standard, which comes from the analog world
> (but the data itself is quite digital, of course). A few pages
> describe how data is transmitted in an analog broadcast (i.e. the line
> 21 waveform). After that, everything else is a description of the 608
> commands.
> 708 describes the new standard which was developed in the digital era.
> 708 cannot be transmitted analogically.

Thanks for the explanations, I will keep it in mind while cataloguing the
samples at hand.

> Yes, pretty much. CCExtractor will be happy to process a file with
> timing + CC data, or even just a small header + CC data.

Unfortunately, there is the "timing +" or "small header +" part that I do
not have. My patch produces packets like that:

00000000: fd8f 2afc 8080 fa00 00fa 0000 fa00 00fa  ..*.............
00000010: 0000 fa00 00fa 0000 fa00 00fa 0000 fa00  ................
00000020: 00fa 0000 fa00 00fa 0000 fa00 00fa 0000  ................
00000030: fa00 00fa 0000 fa00 00fa 0000            ............


If I dump the packets as is in a file (-c copy -f rawvideo), CCExtractor
says "File seems to be an elementary stream, enabling ES mode" but does not
find anything in it. It does find text in the original file, that is not the
problem. And I do not know what header to put on the packets to change
anything to that.

For the record, the command-line I use is this:

./ffprobe_g -show_packets -show_data -select_streams s \
  -f lavfi "movie=subcc_mpeg2.ts[out0+subcc]"

./ffmpeg_g -f lavfi -i "movie=subcc_mpeg2.ts[out0+subcc]" \
  -map s -c copy -f rawvideo -y /tmp/subcc.bin

subcc_mpeg2.ts is the file provided by Anshul in this mail:

-rw-r--r-- 1 cigaes cigaes 7742592 Mar 10 09:10 subcc_mpeg2.ts
310f3748dcd2a9fe652decb2c55f0468792c10de  subcc_mpeg2.ts

The output file I get:

-rw-r--r-- 1 cigaes cigaes 10980 Mar 26 16:49 /tmp/subcc.bin
5a2442a7274c69c43645aab5b7f9735a847417ab  /tmp/subcc.bin

If someone manages to get something useful from that file using my patch,
that would help make things progress.


  Nicolas George
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140326/0f99f7aa/attachment.asc>

More information about the ffmpeg-devel mailing list