2008/5/15 Fan-Cheng Wu ?????\ <fcwu.cs95g at nctu.edu.tw>: > It's no improvement after defining that. :( > Could you explain more about your configuration, and especially what did you do to force PCMU codec to be used and to disable the echo canceller, considering that you use simple_pjsua.c rather than pjsua? The instructions in http://trac.pjsip.org/repos/wiki/audio-check-cpu are for pjsua of course, hence say if you put "--ec-tail 0" argument in simple_pjsua.c most likely it will be ignored. Cheers Benny > 2008/5/14 Nanang Izzuddin <nanang at pjsip.org>: > > > > Try to rebuild without WSOLA implementation by defining: > > #define PJMEDIA_WSOLA_IMP PJMEDIA_WSOLA_IMP_NULL > > in the config_site.h. > > > > WSOLA is currently used for handling audio drift and PLC. > > > > Regards, > > nanang > > > > > > 2008/5/14 Fan-Cheng Wu ?????\ <fcwu.cs95g at nctu.edu.tw>: > >> Thanks for your kindly response, Benny. > >> > >> I tried with PCMU. In my platform, a ARM 9 processor, simple_pjsua is > >> required almost 95% CPU utilization. The only solution seems porting > >> encode/decode to DSP. Does anyone have better suggestions for me?? > >> > >> ( I had complied with http://trac.pjsip.org/repos/wiki/audio-check-cpu > >> to decrease my CPU utlz, but it's useless) > >> > >> -- > >> Fan-Cheng > >> > >> 2008/5/14 Benny Prijono <bennylp at pjsip.org>: > >>> > >>> 2008/5/13 Fan-Cheng Wu ?????\ <fcwu.cs95g at nctu.edu.tw>: > >>> > > >>> > My problem now can be fixed by adding param --ec-tail 0. But, when I was > >>> > talking to pjsua from another user agent which is not pjsua, the "human" > >>> > voice sounded like "robot" voice. It only happened on pjsua with ARM > >>> > platform while Ubuntu is correct. I thinks it's probably the problem on my > >>> > ARM. Anyways, thanks for your help, Nannang. > >>> > > >>> > >>> What codec did you use? If you use high complexity codec then probably > >>> your device couldn't take it. Try with G.711 first. > >>> > >>> > FYI, when I tried "--clock-rate 11025", pjsua would be terminated with > >>> > assertion on ../src/pjmedia/conference.c:1399 > >>> > `cport->rx_buf_count<=cport->rx_buf_cap' > >>> > > >>> > >>> Clock rate 11025 (and 22050) currently can't work without explicit > >>> adjustment to ptime in the source (we need to set ptime to 40 when > >>> clock rate is 11025 and to 20 with clock rate 22050 to make a whole > >>> frame). As Nanang said ticket #532 will fix that, but in the mean time > >>> please don't use clock rate 11025 in any parts of pjmedia. > >>> > >>> I hope these fix your problems, please let me know if they don't. > >>> > >>> Cheers > >>> Benny > >>> > >>> > >>> _______________________________________________ > >>> 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 > >
