[FFmpeg-devel] [PATCH] RDT/Realmedia patches #2

[Ronald S. Bultje]

Hi,

On Wed, Nov 19, 2008 at 5:08 PM, Michael Niedermayer <michaelni at gmx.at> wrote:
> Well, i dont really know your / the RT*P code too well so maybe i
> missunderstand something but AFAICS either
> theres 1 stream per RDTDemuxContext in which case st should be that and
> not the "first of a set"
> or
> there is a set per RDTDemuxContext, in which case the stream index and
> the number of streams in the set seem to be the logic way to keep track

Ah, I see, ok, so it's like this:

General/RTSP/SDP:
- there is one m= SDP line per RTSPStream
- there is one RDT/RTPDemuxContext per RTSPStream

For RTP:
- there is one AVStream per RTPDemuxContext/RTSPStream
- multiple RTPDemuxContext/RTSPStreams can stream the same content
(but possibly in different codecs/bitrates)

For RDT:
- there are (usually) multiple AVStreams per RDTDemuxContext/RTSPStream
- each AVStream within a single RDTDemuxContext/RTSPStream streams the
same content, but possible in different bitrates/codecs
- each RTSPStream/RDTDemuxContext streams unique content compared to
other RTSPStreams/RDTDemuxContexts

So I guess they are semantically not the same, no. I guess I could
still share the "ic" (AVFormatContext) variable between the two,
that's really the only one that I need. So, having cleared that up,
would you prefer a first_stream_in_set variable or would you prefer
just an array of AVStreams in the RDTDemuxContext?

Thanks, and sorry that this is all so complex,
Ronald