[Bug 755125] rtp: RTCP mapping between NTP and RTP time could be capture or send time based

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=755125

--- Comment #5 from Sebastian Dröge (slomo) <slomo at coaxion.net> ---
(In reply to Olivier Crête from comment #3)
> 1) is how I've always seen it done.
> 
> If rtpjitterbuffer has this assumption, I wouldn't be surprised if other
> receivers had the same assumption.
> 
> I'm not sure why don't synchronize the sending of RTCP packets NTP time in
> the SR? Apart from the need to have async=false on the RTCP syncs.

That would mean that we delay all RTCP packets by the pipeline latency... which
kind of defeats the point of RTCP. You want to give feedback ASAP.


Instead of what my patches do, you could also move the NTP timestamp into the
past by the latency to get the same behaviour (just that then the RTP timestamp
in the RTCP packet would be close to what the receiver currently has). But I
think this is not according to the RFC, the RFC explicitly states that the NTP
time should be when the packet is sent.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.