HI, Thanks very much. I will test this tomorrow. Regards, Frederic On Thu, 4 Sep 2008 01:59:08 +0700, "Nanang Izzuddin" <nanang at pjsip.org> wrote: > Hi, > > Sorry, it should be master port, one of pjmedia clock generator, here > is the doc: > http://www.pjsip.org/pjmedia/docs/html/group__PJMEDIA__MASTER__PORT.htm > > There is no default clock rate for wav player, wav player clock rate > follows the wav file clock rate. So if you are using G711, please use > 8000 Hz wav file to avoid resampling. > > Btw, replacing wav player with mem player may increase the > performance, not sure though. > > Regards, > nanang > > > On 04/09/2008, Fr??d??ric CLEMENT <fclement at viatelecom.com> wrote: >> /* > - use master clock (not sound device) for the clock source, since > sound >> device may be disturbed by high CPU load and CPU spikes. > - use clock rate >> 8000 Hz for the master clock (or conf bridge, if you > use it) to avoid >> resampling. >> */ >> >> Thanks very much for this complete answer. >> >> I didn't understand what you mean with 'use master clock'. >> I don't user any hardware but only your wav_player methods. >> The clock rate is 8000 Hz by default, isn't it ? >> >> >> Regards, >> >> Frederic >> >> >> Nanang Izzuddin a ?crit : >> Hi, > > G711 is very low complexity codec >> (see > http://trac.pjsip.org/repos/wiki/PJMEDIA-MIPS, stream >> with G711 in P3 > 700MHz only cost < 0.1% CPU), so the codec shouldn't be the >> problem. > Also your CPU utilization seems to be relatively low (although >> it > supposed to be lower than 30%). So it's not the CPU limitation. > > Please >> see this, in case you >> haven't: > https://trac.pjsip.org/repos/wiki/FAQ#high-perf. > And >> also please try this, in case you haven't: > - use master clock (not sound >> device) for the clock source, since > sound device may be disturbed by high >> CPU load and CPU spikes. > - use clock rate 8000 Hz for the master clock (or >> conf bridge, if you > use it) to avoid resampling. > > Btw, is there any chance >> that the problem is in the caller/tester > (e.g: the app tester can only >> handle 20 concurrent calls, so you > should run 2 instances of app tester on 2 >> different machines)? > > Regards, > nanang > > > On 03/09/2008, Fr??d??ric CLEMENT >> <fclement at viatelecom.com> wrote: > >> My Load average is about 1.00 (34 % CPU utilization). > The codec I am using >> is Alaw (g711). > Can I expect to handle 100 calls with a Core2Duo or QuadCore >> Xeon ? > > Should I user another codec ? My switch have many implemented. >> (Dialogic > Cantata IMG). > > Regards, > > Frederic > > > Nanang Izzuddin a ?crit : > Hi, >> Which codec are you using? How about the CPU utilization? Since > >> it's > >> may be caused by CPU limitation in doing media processing > >> (especially > >> encoding). > > Regards, > nanang > > > On 03/09/2008, Fr??d??ric CLEMENT > >> <fclement at viatelecom.com> wrote: > >> >> Hi again all, > >> I am still building an IVR server and testing the max > >> possible account of > >> calls that can be handled by pjsip. > I set > >> PJSUA_MAX_CALLS to 100 > >> The value inf config_site.h for > >> PJMEDIA_SOUND_BUFFER_COUNT is set to 16. > >> My application just take calls and > >> play a sound file looped. I use the pjsua > >> high level methods > >> (pjsua_player_create ....) > >> The machine is a Pentium IV 1.9 Ghz with enough > >> memory. > >> When I send 20 calls, the sound stays being correct. > > If I send 30 > >> simultaneous calls, I get the same result than described on > >> your site > >> : > >> http://trac.pjsip.org/repos/wiki/audio-problem-dropouts > > The >> sound I have is the same than the stutter.wav found > there. > >> > http://trac.pjsip.org/repos/attachment/wiki/audio-problem-dropouts/stutter.wav > > Wich >> optimisations can be done ? My goal is about 100 simultaneous > calls. > >> Regards, > > Frederic > > (Sorry again for my very bad > >> english) > >> _______________________________________________ > Visit > >> our blog: http://blog.pjsip.org > >> pjsip mailing > >> list > >> pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > >> _______________________________________________ > >> Visit our > >> blog: http://blog.pjsip.org > >> pjsip mailing > >> list > >> pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> _______________________________________________ > Visit our >> blog: http://blog.pjsip.org > > pjsip mailing >> list > pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > > > >> _______________________________________________ > Visit our >> blog: http://blog.pjsip.org > > pjsip mailing >> list > pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip at lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip at lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
