Hi Tanger,
Sorry, I forgot to mention sndtest in 0.8.0 is actually buggy, it
always reports no drift detected. It has been fixed months ago in the
SVN trunk. So please check the drift with the latest SVN.
About audio problem, please make sure whether it is caused by AEC (by
comparing ec-tail > 0 to ec-tail = 0) and please try with some sound
device clock rates especially the clock rate with less drift.
FYI, in the latest SVN, pjsua has new option --snd-clock-rate, to
enable different clock rate setting between sound device and
conference bridge.
Regards,
nanang
On 17/04/2008, cat tanger <tanger_wendy at yahoo.com.cn> wrote:
>
>
>
> Hi ,Nanang Izzuddin.
> thank you very much.
>
> 1. I test my sound cards do not clock drift.
> 14:45:49.109 os_core_win32. pjlib 0.8.0 for win32 initialized
> 14:45:49.187 pasound.c PA message: Pa_GetDeviceInfo: Num input
> channels reported as 65535! Changed to 2.
> 14:45:49.187 pasound.c PA message: Pa_GetDeviceInfo: Num
> output channels reported as 65535! Changed to 2.
> 14:45:49.343 pasound.c PortAudio sound library initialized,
> status=0
> 14:45:49.343 pasound.c PortAudio host api count=3
> 14:45:49.343 pasound.c Sound device count=8
> 14:45:49.359 pjlib select() I/O Queue created (00A95CEC)
> 14:45:49.359 sndtest.c Found 8 devices:
> 14:45:49.359 sndtest.c 0: Microsoft ????? - Input (capture=2,
> playback=0)
> 14:45:49.359 sndtest.c 1: Realtek HD Audio Input (capture=2,
> playback=0)
> 14:45:49.359 sndtest.c 2: Microsoft ????? - Output
> (capture=0, playback=2)
> 14:45:49.359 sndtest.c 3: Realtek HD Audio output (capture=0,
> playback=2)
> 14:45:49.359 sndtest.c 4: ????????? (capture=2, playback=0)
> 14:45:49.359 sndtest.c 5: Realtek HD Audio Input (capture=2,
> playback=0)
> 14:45:49.359 sndtest.c 6: ??????? (capture=0, playback=2)
> 14:45:49.359 sndtest.c 7: Realtek HD Audio output (capture=0,
> playback=2)
> 14:45:49.375 sndtest.c Testing playback device Microsoft ?????
> - Output
> 14:45:49.375 sndtest.c Testing capture device Microsoft ?????
> - Input
> 14:45:49.593 sndtest.c Please wait while test is in progress
> (~11 secs)..
> 14:46:00.718 sndtest.c Dumping results:
> 14:46:00.718 sndtest.c Parameters: clock rate=16000Hz, 160
> samples/frame
> 14:46:00.718 sndtest.c Playback stream report:
> 14:46:00.718 sndtest.c Duration: 10s.000
> 14:46:00.718 sndtest.c Frame interval: min=7.852ms,
> max=12.159ms
> 14:46:00.718 sndtest.c Jitter: min=1.780ms, avg=3.541ms,
> max=4.307ms
> 14:46:00.734 sndtest.c Capture stream report:
> 14:46:00.734 sndtest.c Duration: 10s.010
> 14:46:00.734 sndtest.c Frame interval: min=0.022ms,
> max=24.051ms
> 14:46:00.734 sndtest.c Jitter: min=9.965ms, avg=21.575ms,
> max=24.013ms
> 14:46:00.734 sndtest.c Checking for clock drifts:
> 14:46:00.734 sndtest.c No clock drifts is detected
> 14:46:00.750 sndtest.c Test completed with some warnings
>
> 2. i say aec work well mean aec work in ptime=10 mode can cancel partial
> echo , tail_len =200.
> but audio not good , with clutter.
> the attichment are aec_out.sw,aec_rec.sw and aec_play.sw files.
> clock_rate =16000 ,bits_per_sample = 16,channel_cnt =1.
>
> 3. it runing on vista easy result aec reset , and often happen on the end of
> far user burst speak.
> reset times too faster, 20ms reset 15 times , and can't hear what the
> far user say.
> 16:32:33.808 aec02FC94E0 AEC reset, delay=-9284, prefetch=10
> 16:32:33.809 aec02FC94E0 AEC reset, delay=-9285, prefetch=10
> 16:32:33.841 aec02FC94E0 AEC reset, delay=-9286, prefetch=10
> 16:32:33.844 aec02FC94E0 AEC reset, delay=-9287, prefetch=10
> 16:32:33.873 aec02FC94E0 AEC reset, delay=-9288, prefetch=10
> 16:32:33.874 aec02FC94E0 AEC reset, delay=-9289, prefetch=10
> 16:32:33.876 aec02FC94E0 AEC reset, delay=-9290, prefetch=10
> 16:32:33.908 aec02FC94E0 AEC reset, delay=-9291, prefetch=10
> 16:32:33.909 aec02FC94E0 AEC reset, delay=-9292, prefetch=10
> 16:32:33.940 aec02FC94E0 AEC reset, delay=-9293, prefetch=10
> 16:32:33.941 aec02FC94E0 AEC reset, delay=-9294, prefetch=10
> 16:32:33.972 aec02FC94E0 AEC reset, delay=-9295, prefetch=10
> 16:32:33.973 aec02FC94E0 AEC reset, delay=-9296, prefetch=10
> 16:32:34.001 aec02FC94E0 AEC reset, delay=-9297, prefetch=10
> 16:32:34.003 aec02FC94E0 AEC reset, delay=-9298, prefetch=10
> 16:32:34.004 aec02FC94E0 AEC reset, delay=-9299, prefetch=10
> 16:32:34.034 aec02FC94E0 AEC reset, delay=-9300, prefetch=10
> 16:32:34.036 aec02FC94E0 AEC reset, delay=-9301, prefetch=10
> 16:32:34.069 aec02FC94E0 AEC reset, delay=-9302, prefetch=10
> 16:32:34.070 aec02FC94E0 AEC reset, delay=-9303, prefetch=10
> 16:32:34.100 aec02FC94E0 AEC reset, delay=-9304, prefetch=10
> 16:32:34.101 aec02FC94E0 AEC reset, delay=-9305, prefetch=10
> 16:32:34.129 aec02FC94E0 AEC reset, delay=-9306, prefetch=10
> 16:32:34.130 aec02FC94E0 AEC reset, delay=-9307, prefetch=10
> 16:32:34.161 aec02FC94E0 AEC reset, delay=-9308, prefetch=10
> 16:32:34.164 aec02FC94E0 AEC reset, delay=-9309, prefetch=10
> 16:32:34.165 aec02FC94E0 AEC reset, delay=-9310, prefetch=10
> 16:32:34.196 aec02FC94E0 AEC reset, delay=-9311, prefetch=10
> 16:32:34.197 aec02FC94E0 AEC reset, delay=-9312, prefetch=10
>
> regards.
>
> tanger.
>
>
>
>
> Date: Thu, 17 Apr 2008 11:44:25 +0700
> From: "Nanang Izzuddin" <nanang@xxxxxxxxx>
> Subject: Re: about audio question!
> To: "pjsip list" <pjsip at lists.pjsip.org>
> Message-ID:
>
> <c7f43120804162144u1fd725c6w3e4a4fea4fdcbb7 at mail.gmail.com>
> Content-Type: text/plain; charset=GB2312
>
>
> Hi Tanger,
>
> Some sound cards do have clock drift, recording and playing do not
> have exact same clock rate, (and this usually will hurt the AEC, but
> not in your case?). sndtest sample app can be used to check drift
> level of your sound card.
>
> However it is a bit strange on your case, since usually AEC will work
> well when there is no clock drift, and AEC becomes inefficient/not
> working when there is drift. But, your experience is on the contrary
> of those usual behavior.
>
> So, if you are sure that the AEC works well (echo is canceled) in
> ptime==10, where the drift occurred, then just use this setting :)
>
> You may want to try opening the sound card in clock rate 48kHz which
> usually produce less drift.
>
>
> Regards,
> nanang
>
>
> On 17/04/2008, cat tanger <tanger_wendy at yahoo.com.cn> wrote:
> >
> >
> >
> > hi,all,
> > please help me!
> > I use pjproject-0.8.0,
> >
> > when i set audio_frame_ptime = 20,aec = 1, then PaRecorderCallback and
> > PaPlayerCallback function run exactly, but aec can't work well, do not
> > cancell echo .
> >
> > when i set audio_frame_ptime = 10,aec = 1, then aec can work well . but
> > the PaRecorderCallback and PaPlayerCallback function do not work well,
> > portaudio callback the PaRecorderCallback times far more callback
> > PaPlayerCallback times, it result the aec continue reset. i find underflow
> > in the PaPlayerCallback.
> >
> > pls help me!
> > thanks!
> >
> >
> > test code:
> >
> > static int capcounts = 0,playcounts = 0;
> >
> > static int PaRecorderCallback(const void *input,
> > void *output,
> > unsigned long frameCount,
> > const PaStreamCallbackTimeInfo* timeInfo,
> > PaStreamCallbackFlags statusFlags,
> > void *userData )
> > {
> > .....
> > capcounts ++;
> > if (capcounts%500 == 0)
> > {
> > PJ_LOG(5,(THIS_FILE, "Recorder thread capcounts = %d,playcounts =
> > %d",capcounts,playcounts));
> > }
> > ....
> > return paAbort;
> > }
> >
> > static int PaPlayerCallback( const void *input,
> > void *output,
> > unsigned long frameCount,
> > const PaStreamCallbackTimeInfo* timeInfo,
> > PaStreamCallbackFlags statusFlags,
> > void *userData )
> > {
> > ....
> > playcounts ++;
> > if (playcounts%500 == 0)
> > {
> > PJ_LOG(5,(THIS_FILE, "Player thread capcounts = %d,playcounts = %
> > d",capcounts,playcounts));
> > }
> > ....
> > return paAbort;
> > }
> >
> >
> > print :
> > 11:11:06.703 aec00EA2B70 AEC reset, delay=-65, prefetch=10
> > 11:11:06.921 pasound.c Recorder thread capcounts = 500,playcounts =
> > 434
> > 11:11:07.062 aec00EA2B70 AEC reset, delay=-66, prefetch=10
> > 11:11:07.609 pasound.c Player thread capcounts = 568,playcounts =
> 500
> > 11:11:07.734 aec00EA2B70 AEC reset, delay=-69, prefetch=10
> > 11:11:11.921 pasound.c Recorder thread capcounts = 1000,playcounts =
> > 903
> > 11:11:11.921 aec00EA2B70 AEC reset, delay=-89, prefetch=10
> > 11:11:12.828 pasound.c Player thread capcounts = 1089,playcounts =
> > 1000
> > 11:11:13.140 aec00EA2B70 AEC reset, delay=-90, prefetch=10
> > 11:11:16.906 pasound.c Recorder thread capcounts = 1500,playcounts =
> > 1376
> > 11:11:17.203 aec00EA2B70 AEC reset, delay=-123, prefetch=10 1
> > 11:11:18.390 pasound.c Player thread capcounts = 1646,playcounts =
> > 1500
> > 11:11:18.390 aec00EA2B70 AEC reset, delay=-140, prefetch=10
> > 11:11:21.921 pasound.c Recorder thread capcounts = 2000,playcounts =
> > 1817
> > 11:11:23.171 aec00EA2B70 AEC reset, delay=-183, prefetch=10
> > 11:11:23.828 pasound.c Player thread capcounts = 2191,playcounts =
> > 2000
> > 11:11:24.234 aec00EA2B70 AEC reset, delay=-192, prefetch=10
> > 11:11:26.921 pasound.c Recorder thread capcounts = 2500,playcounts =
> > 2253
> > 11:11:26.937 aec00EA2B70 AEC reset, delay=-239, prefetch=10 1
> > 11:11:29.703 pasound.c Player thread capcounts = 2779,playcounts =
> > 2500
> > 11:11:29.937 aec00EA2B70 AEC reset, delay=-278, prefetch=10
> > 11:11:31.906 pasound.c Recorder thread capcounts = 3000,playcounts =
> > 2681
> > 11:11:32.140 aec00EA2B70 AEC reset, delay=-311, prefetch=10
> >
> >
>
> ________________________________
> ????????????
> _______________________________________________
> 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
>
>
>
