On 2/18/06, Ian Service <[EMAIL PROTECTED]> wrote: > IRC seems to be useless for me on this one, so I figured I'd try you guys > before the bugs.digium folk. I'm running latest SVN on two asterisk > machines.
DTMF, my old enemy. I've done extensive testing with RFC2833 DTMF over the last several months, and it seems to be everyones fault (Asterisk and Cisco in my case, but I'm sure other devices as well). I'll give a very brief overview here, but I encourage you to read, and if appropriate, comment on the bugs already open as opposed to opening a new bug. They can be found here: http://bugs.digium.com/view.php?id=5970 http://bugs.digium.com/view.php?id=6027 5970 is the more appropriate bug to comment on as it has much more information about the real issue with RFC2833 in Asterisk. Basically, when you send DTMF via RFC2833 stylez, each digit pressed will send 6 packets -- 3 start events and 3 end events. If the far end recieves the 3 start events in order, followed by 3 end events, then there is no problem and DTMF is completed successfully. However, if the events arrive out of order, then you end up with the far end detecting double and triple DTMF (yes, this is a problem in Asterisk as well (at least up to 1.2.1 where I did my testing)). One of the problems is that Asterisk sends all 6 events almost instantaneously, which over the public Internet is likely to arrive out of order due to jitter and latency. There was an idea that placing some sort of delay between the packets (and at the very least between the start and end transition events) might help resolve the problem, but that seems pretty hackish. Ideas welcome :) -- Leif Madsen. http://www.leifmadsen.com http://www.oreilly.com/catalog/asterisk
