Yes, the dlsr computation code was broken, I've fixed it in CVS, but I haven't tested it yet, so be careful. Also, consider that ccRTP follows RFC 3550, which adds some timing rules not specified in RFC 1889.
Thanks for spotting these bugs! On Tue, Apr 19, 2005 at 12:35:38PM +0900, Takegami, Takashi wrote: > Hello, > > I don't understand DLSR in ReceiverReport. > In "control.cpp", DLSR is made from gettimeofday() and NTP > timestamp(in SenderReport) as (gettimeofday - NTP) . > Isn't this wrong ? > > DLSR is the time between receiving the last SR and sending this RR. > so, DLSR = gettimeofday() - the time when receiving the last SR. > (see fig. RFC1889 section6.3.2) > > so that, I cannot calculate Round-Trip delay by using ccRTP... > > > Hope this helps, > -- > Takegami > > > _______________________________________________ > Ccrtp-devel mailing list > [email protected] > http://lists.gnu.org/mailman/listinfo/ccrtp-devel _______________________________________________ Ccrtp-devel mailing list [email protected] http://lists.gnu.org/mailman/listinfo/ccrtp-devel
