May be...
I also thought that it may depend on some parameter like Jitter sent in the
RTCP packets up to that time. It may also be some kind of protection against
packet losses occured for the last ('t'-'t1') seconds. But what I am mostly
interested is if this number is a determinstic number, which can also be
calculated in the client side, so that the client will also know this 't1'
beforehand.
BR
Emre
> -----Original Message-----
> From: EXT Francesco Schiavon [mailto:email@hidden]
> Sent: 23. November 2000 23:14
> To: email@hidden; email@hidden
> Subject: Re: Starting timestamps of packets after a Pause and a Play
> reques t
>
>
> On 11/23/00 13:05 email@hidden wrote:
>
> >What I noticed from the network traffic monitoring is that,
> even the new
> >PLAY request asks for the movie to start at time, let's say
> 't' (in the RTSP
> >PLAY request 'Range: npt = t - '), Darwin starts sending
> packets belonging
> >to a time 't1' where t1 < t. I made some tests and this 't1'
> seems to be
> >random, ranging from 0.5 to 1 second (or may be even more?).
> Does anyone
> >know how this new transmission starting time is calculated
> in Darwin Server?
> >Why doesn't it just start sending packets belonging to
> requested start time
> >'t'?
>
> I'm just guessing here, but would it have to do with the fact
> that it has
> to buffer? It sounds like a "pre-roll" idea to me.
>
>
> -
>
> francesco schiavon
>
> vancouver film school - new media
> multimedia.edu
> vfs.com
