On Thu, Apr 3, 2008 at 10:58 PM, Norman Franke <norman at myasd.com> wrote: > > Unfortunately, that's going to be hard. I can let our folks use this > normally since it makes some calls impossible to hear, but not all. I can > try to reproduce it on my end, though. As someone mentioned, it may have > something to do with loud audio. Not sure. I'll see what I can find out. > A little bit more info. We did two quite major fixes in the last few weeks. First is related to audio buffering to cater audio burst in sound port, the bridge, and splitcomb (using something called delaybuf, which utilizes another thing called wsola, all of them are new). We don't think this would cause the stutter, and in fact the purpose of this fix is to remove the stutter altogether! But there is a second fix that we did that may cause the stutter, that is the modification in the jitter buffer to improve the delay (ticket http://trac.pjsip.org/repos/ticket/505, which I just re-opened). If this is the reason, then the log should display something about jitter buffer operation when the stutter happens. If it's too hard to get a log, maybe you can rollback the jitter buffer to the revision before ticket #505 and see if this is really the culprit. Thanks Benny
