[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [AVT] RE: Carrying SMPTE time-codes in RTP streams, discussion email



At 1:25 PM +0000 2/26/05, Colin Perkins wrote:
On 24 Feb 2005, at 08:47, Dave Singer wrote:
At 1:49 PM -0800 2/23/05, lazzaro wrote:
[Magnus and Dave in discussion on SMPTE]

A. Is it acceptable to get the update a bit late? Using RTCP can easily result in that the time code update arrives up to a few seconds late. However if some latency is acceptable, the RTCP bandwidth and the usage of the AVPF profile can ensure that one maintain a limit on the latency to send the update.


I *think* so...

This is an important point, so further feedback on this is crucial.

I think the right answer to the question is "depends on the app", and separates into two questions:

A. For the app, is associating the wrong SMPTE code with the RTP
stream for a few seconds a disaster?

B. For the app, how much advance notice does it have of an upcoming
SMPTE change?

Apps where A="yes" and B="very little if any" are probably not going
to be happy with the RTCP approach ... apps where A="no" will be
fine with the RTCP approach ... the final class of apps is "it depends".

An example of "B=little if any" is when the SMPTE is changing because
an operator is hitting buttons on a tape-deck controller, and is expecting
the RTP-controlled rack-mount to respond with 5 ms latency over a local
LAN in a reliable way.

Note that the mapping itself in the RTCP can always be right; it's merely a question of whether it arrives too late for the terminal to be able to apply it to the media from the identified timestamp. If indeed the client wants a short buffer, and the source gets no advance notice, and it's not possible to send immediate RTCP packets, there may be an issue.

We might be able to use RTP/AVPF to send early RTCP feedback? Or there might be non-RTP signalling which can convey the mapping (e.g. via RTSP)?



These are sender reports, not receiver. But yes, RTSP could say what the initial mapping was also (just like the NTP association sent now).
--
David Singer
Apple Computer/QuickTime


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