[FFmpeg-soc] RTP packetizers

[Martin Storsjö]

On Mon, 9 Aug 2010, Luca Barbato wrote:

> On 08/09/2010 01:39 PM, Martin Storsjö wrote:
> > As far as I know, a theora packet is a complete frame. I guess Luca B can 
> > shed some light on this issue?
> 
> the terminology clash had been fixed in vorbis rfc while the theora rfc
> got dust piling up, you should pick the vorbis rfc and change vorbis
> with theora and act upon it. a theora frame can be fragmented in many
> rtp packets or many theora smaller frames could be put in a single rtp
> packet.

Any explanation to the RFC recommendation that you should put more than 
one frame into one packet, then? Since RTP isn't strictly fixed frame 
rate, the receiver can't properly set timestamps for the later frames in 
the packet, right? Unless it tries to guess the normal frame rate from 
earlier packets...

// Mrtin