Re: [AVT] Vorbis RTP issues list for Vienna

> Magnus Westerlund <magnus.westerlund@ericsson.com> writes:
>
> I think that we need to clear consider how to get these codebooks
> transported. A larger than 64kb codebook should neither be transported
> in SDP or in RTCP. I think that using HTTP can clearly be an
> alternative, and only provide the URI to the codebook.
>
> I wouldn't try to transport a codebook over RTCP that results in a
> RTCP packet many time the MTU due to the high number of IP
> fragments. Further such a packet will severely reduce the RTCP report
> rate in both directions.

I may be incorrect here, but from my discussions with Phil earlier
in the year, the codebook issue has these requirements:

 -- May change every few minutes (say, once per song when used for 
    Internet radio apps).
 -- Tightly synchronized with the stream: if you don't have the new
    codebook in time for the first packet that needs it, its a 
    catastrophe.
 -- Reliable reception a must, no built in resiliency if it is
    sent as a stream of UDP packets and some but not all arrive.

One way to handle this is to have two synchronized streams in the
media session: a UDP RTP/AVP stream for the media data, and a
TCP stream (perhaps an RTP/AVP stream if a solution like
draft-lazzaro-avt-rtp-framing-contrans happens, or perhaps an
opaque TCP stream) for the codebooks.  In either case, the
synchronization would happen as lip-sync synchronization happens.

-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu     www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------

_______________________________________________
Audio/Video Transport Working Group
avt@ietf.org
https://www1.ietf.org/mailman/listinfo/avt