In article <[EMAIL PROTECTED]>, BJ Weschke <[EMAIL PROTECTED]> wrote: > What version of * are you running? There was a bug that was posted a > few weeks back where when not using the "q" option it was possible for > legs of the conference to get further separated from each other > (sometimes up to 3-5 secs after 60 mins of conferencing). The bug is > in mantis and there is a patch that seems to keep that down to no more > than 1/2 to 3/4 second after several hours of conferencing, but I'm > not certain that this patch made it back into the CVS-HEAD branch.
It didn't (I was the proponent of the patch). I disagreed with Mark's alternative solution, and still do, so it's sort of stalled for now. The patch plays tones and announcements destined for the whole conf using a separate thread. Playing and waiting for them inline causes the channel's own queue not to be serviced, developing a backlog. This applies mainly to enter/leave sounds and '"Fred" has joined' announcements. I believe there are other lag issues in non-Zap meetme that are not addressed by my patch, but I haven't had time to investigate them thoroughly yet. If I set up a meetme between two SIP phones with the 'q' flag set, I start off with negligible delay. If I just leave the conference running, that delay builds, up to the 1/2 to 3/4 of a second as reported above, but then sometimes disappears before building again. I haven't yet determined whether the lag is on the SIP->conf leg or the conf->SIP leg, or both. My suspicion is that it is drift due to clock differences between the ZAP timer and the SIP device, but am not sure how to determine when a frame should be dropped or duplicated. Cheers Tony -- Tony Mountifield Work: [EMAIL PROTECTED] - http://www.softins.co.uk Play: [EMAIL PROTECTED] - http://tony.mountifield.org _______________________________________________ Asterisk-Users mailing list [email protected] http://lists.digium.com/mailman/listinfo/asterisk-users To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-users
