[FFmpeg-devel] theora depayloader

Martin Storsjö martin
Fri Mar 19 15:19:56 CET 2010


On Fri, 19 Mar 2010, Ronald S. Bultje wrote:

> On Fri, Mar 19, 2010 at 5:23 AM, Martin Storsj? <martin at martin.st> wrote:
> >> + ? ?case CODEC_ID_THEORA:
> >> + ? ? ? ?ff_theora_parse_fmtp_config(codec, ctx, attr, value);
> >> ? ? ?case CODEC_ID_VORBIS:
> >> ? ? ? ? ?ff_vorbis_parse_fmtp_config(codec, ctx, attr, value);
> >> ? ? ? ? ?break;
> >
> > Can this be done through parse_sdp_a_line as for e.g. AMR? Then we
> > wouldn't need to hardcode these format specific parsing routines here...
> > Yes, this is the way it's done for vorbis, but I think it would be better
> > to migrate new things to use RTPDynamicProtocolHandler properly instead of
> > hardcoding these. (A patch for fixing this for vorbis would be welcome,
> > too, unless Ronald or someone disagrees...)
> 
> Missing break in the above code, btw. I wonder how it didn't segfault.

Ah, missed that one...

> We don't use parse_sdp_a_line() because dynamic payloads are set up
> after this SDP line comes in. It's a bit ugly, it's worth a FIXME but
> we decided to not care about it when the Vorbis depayloader was
> accepted.

Ok then.

Why is that, btw? Does the Theora/Vorbis SDP files contain a=fmtp before 
the corresponding a=rtpmap, or why is this any different from other 
formats having fmtp lines?

// Martin



More information about the ffmpeg-devel mailing list