Hello Nanang, So many Thanks for your detail Information. My problem is now solved. Actually it was possible because of your detail information. I only disable VAD on pjsip. Now I am not geting any break. Thanks Salahuddin On Mon, Aug 4, 2008 at 3:38 PM, Nanang Izzuddin <nanang at pjsip.org> wrote: > Hi Salahuddin, > > Thanks so much for the input. I think calling PlayCb() after RecCb() > may be a good alternative for CPU idle/normal condition, since in this > condition it seems there is no clock drift/skew between mic & spk (at > least in E65) so the delay should be kept constant and no breaking of > underflow. However I experienced temporary clock drift when there is a > CPU load spike (like printing log in console, in a pjsua call using > speex) and it took few seconds to recover the synchronized clock rate > (clock drift symptom is analyzed by observing delay buffer's log > messages, i.e: periodically generating/reducing samples), this > temporary drift may result either underflow/breaking or the delay > increasing badly (I read somewhere that the APS message queue size is > 100, so the speaker delay can reach max 100 frames == 2 seconds for > 20ms ptime). > > Actually I was looking for discussion topics or articles about > where/when to call PlayCb() 'correctly' in APS, but still no luck. > After a few experiments (I've also desperately tried PlayCb controlled > by RTimer), I decided to put the PlayCb() there (in the > EPlayCommQueue) since the voice result was good and it seems to be > safe when clock drift occurs. > > However, your case is still a mystery why the voice was breaking ONLY > when the mic receive no sound/silence. Perhaps you can analyze on the > frame size and type (the first 2 bytes) from mic when silence, or how > many times the RecCb and PlayCb was called in silence (they should be > same or a small diff). > > Regards, > nanang > > > 2008/8/1 Salahuddin Ahmed <bd.rubel at gmail.com>: >> Hi Nanang, >> Thanks For your reply. I am sorry for late reply. Actually I am not >> so much matured in codec area. >> Breaking is not 10 times in a second. it is 1/2 times in a second but >> the length of breaking duration is more. >> I make loopback(mic->spk) test this is alright(getting no break). I >> make another wild test :) I call play_cb after recb_cb and comment >> play_cb from EPlayCommQueue. This make little bit delay but the >> breaking reduce much more. I guess it is not the perfect possition to >> call play_cb. Another thing When I make continuous sound in mic then I >> didnt get any break in play stream. >> >> My and ur Environment are same. I use E65, 3rd Ed MR SDK, APS 2.43. >> >> Thanks, >> Salahuddin >> >> On Thu, Jul 31, 2008 at 8:04 PM, Nanang Izzuddin <nanang at pjsip.org> wrote: >>> Hi Salahuddin, >>> >>> How frequent is the breaking noise (e.g: about ten times in a second)? >>> Or completely noise? >>> >>> It can be the frames from mic are normal G.711 frames when not >>> silence, but it returns non-normal frames (e.g: CNG frames or perhaps >>> lower rates) when silence (non-normal frames are not handled by now). >>> However this isn't supposed to happen since VAD and CNG are disabled. >>> Could you specify your environment (e.g: application, device, SDK >>> version, APS version)? Was that in a call or looped back mic->spk? I >>> will try to build a 'similar' environment if possible and start >>> 'debugging' around, since the problem doesn't occur in my current >>> environment (symsndtest/symbian_ua/symbian_ua_gui, E65, S60 3rd ed MR >>> SDK, APS 2.43). >>> >>> To be honest, I don't have a lot of experience with APS and know >>> almost nothing on APS issues/tricks. So opinions or contributions on >>> the APS integration are greatly appreciated. >>> >>> Regards, >>> nanang >>> >>> >>> 2008/7/31 Salahuddin Ahmed <bd.rubel at gmail.com>: >>>> Hello Nanang, >>>> >>>> Firstly so many thanks for your APS integration. I can successfully >>>> install and use it But I got an problem... The play is not continuous. >>>> It contain so many breaks if I dont say anything in mic. If I make >>>> some sound in mic then the play will continuous. I can't understand >>>> what is problem. >>>> >>>> thanks >>>> >>>> On Wed, Jul 30, 2008 at 7:13 PM, Nanang Izzuddin <nanang at pjsip.org> wrote: >>>>> Hi Karthik, >>>>> >>>>> Was the problem in running or installation? Any error message/code >>>>> issued? The APS code is actually experimental, however it seems to >>>>> work smoothly on E65. Please make sure you have the APS server >>>>> installed on your device and use the correct target device of SDK API >>>>> Plug-In. Please also see http://trac.pjsip.org/repos/wiki/APS. >>>>> Feedbacks on this are very welcomed. >>>>> >>>>> Btw, could you share some hints regarding Audio Routing API >>>>> integration here? May it be useful for the others. >>>>> >>>>> Thanks & regards, >>>>> nanang >>>>> >>>>> >>>>> 2008/7/30 Karthik Babu <cytrion at gmail.com>: >>>>>> Hello Nannang , >>>>>> >>>>>> I am able to use the Audio Routing API successfully on my n95 , but this is >>>>>> limited to fp1 and fp2 phones . >>>>>> >>>>>> So I planned to use APS (for non fp1 and fp2 devices) and I did notice >>>>>> that the APS is available in the current svn trunk . I encounetered few >>>>>> issues with this but could finally make a build. But the exe fails on my N95 >>>>>> . >>>>>> >>>>>> Can you please advice ? >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Karthik >>>>>> >>>>>> _______________________________________________ >>>>>> 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 >>>>> >>>> >>>> >>>> >>>> -- >>>> Salahuddin Ahmed >>>> Software Engineer >>>> Genuity Systems Ltd. >>>> www.genuitysystems.com >>>> Tel. 88-02-8057038-9, 88-02-8079997 >>>> Sip address: sip:86233 at iptel.org >>>> Skype : bdrubel >>>> LinkedIn: www.linkedin.com/in/salahuddin >>>> >>>> _______________________________________________ >>>> 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 >>> >> >> >> >> -- >> Salahuddin Ahmed >> Software Engineer >> Genuity Systems Ltd. >> www.genuitysystems.com >> Tel. 88-02-8057038-9, 88-02-8079997 >> Sip address: sip:86233 at iptel.org >> Skype : bdrubel >> LinkedIn: www.linkedin.com/in/salahuddin >> >> _______________________________________________ >> 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 > -- Salahuddin Ahmed
