On Tue, 12 Oct 2004, Christopher L. Wade wrote: > Darren Sessions wrote: > > Maybe I'm not familiar enough with the internals of Asterisk to > > understand what kind of timing you're after. I assumed you were just > > after a reference clock. > > > > AFAIK, asterisk needs an 'interrupt generator' basically. Something to > wake up a sleeping thread on a _timed_ basis to let to do its job > consistently. Otherwise you'll get clicks/pops/audio-dropouts, etc.
Rtp needs timing to allow it to schedult outgoing packets at regular time intervals. For this the period is something like 20ms and the timing needs to be +/- a few ms. The more precis the outgoing packets are scheduled the less jitter is introduced by the sending. Another timing problem is the timing needed by the zaptel kernel driver. It calculates the sums for zaptel conferences (used by meetme) every ms. This needs to be scheduled every ms. I guess the mixing is done so often to keep the latency of the tdm path low. The timing comes from the various zaptel cards normally. Modern linux kernels can provide this as well. A third kind of timing is for the physical frames on isdn cards. These can be related within a card (i.e. one span is used as the source for another span on the same card), but not between cards. Also, I don't think the linux kernel can ever provide the clock source. If it could, different card could be run off the same clock source. Peter _______________________________________________ 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
