Hi All, I experienced pjsua on TI Davinci, which configured ./aconfigure --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG -mcpu=arm926ej-s' --disable-floating-point and defined "#define PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. When I run pjsua to record sound of 10 second from mic to a file, the file would be extended to 20 second long and sound lower pitch than original one. (Play it with double rate, it sound fine.) I suppose it's because recording device has wrong recording rate. Anybody knows how can I adjust it? p.s. If I recorded a default ring sound from port 02 (Port #02[ 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the problem come from recording device (?). The following are the steps I recorded sound from mic. ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav --clock-rate=8000 --snd-clock-rate=8000 ... >>> cc 0 1 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open devices @8000 Hz 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, samples per frame=160, tail length=200 ms, latency=96 ms 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) Best Regards, FCWu

Hi, Not sure what's going on, but I agree it seems to be a sound device issue, e.g: it may be actually opened in either stereo or 16KHz. I found this old post that may be related to this issue: http://mlblog.osdir.com/linux.davinci/2006-10/msg00283.shtml, as it says "Davinci OSS driver don't work correctly in mono mode and driver still return two samples per tick". Perhaps you can check this by running pjsua in stereo mode (with param '--stereo') and see the mic recording result. Then you can also check pjsua with clock rate 16KHz. In case it *is* sound device issue, you can also search/post a question in the PortAudio forum (PJSIP uses PortAudio on linux platform). Moreover, you can implement your own sound device wrapper, please see http://trac.pjsip.org/repos/wiki/External_Sound_Device for the build config. Regards, nanang On Thu, Oct 9, 2008 at 1:25 PM, Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw> wrote: > Hi All, > > I experienced pjsua on TI Davinci, which configured ./aconfigure > --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG > -mcpu=arm926ej-s' --disable-floating-point and defined "#define > PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. > When I run pjsua to record sound of 10 second from mic to a file, the > file would be extended to 20 second long and sound lower pitch than > original one. (Play it with double rate, it sound fine.) I suppose > it's because recording device has wrong recording rate. Anybody knows > how can I adjust it? > > p.s. If I recorded a default ring sound from port 02 (Port #02[ > 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the > problem come from recording device (?). > > The following are the steps I recorded sound from mic. > ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav > --clock-rate=8000 --snd-clock-rate=8000 > ... >>>> cc 0 1 > 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open > devices @8000 Hz > 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, > samples per frame=160, tail length=200 ms, latency=96 ms > 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) > > > > Best Regards, > FCWu > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >

Hi, I upload two wav files to explain my problem. The file recorded by this way(http://trac.pjsip.org/repos/wiki/audio-how-to-record) could be found here http://moon.cse.yzu.edu.tw/~s912356/output_people_8000.wav. It sound slow and low pitch. The correct one should sound like this http://moon.cse.yzu.edu.tw/~s912356/output_people_16000.wav. I made following testings to clarify my problem. 1. "cat /dev/dsp > file", and then "cat file > /dev/dsp" => Correct 2. Record and Play audio by portaudio(testing src is third_party/portaudio/test/patest_record.c) => Correct 3. PJSUA Looping-back testing(http://trac.pjsip.org/repos/wiki/audio-check-loopback) => Correct 4. Record any audio either from file or mic by PJSUA, and Play them on Davinci by PJSUA and sample/playfile => Correct 5. Record any audio either from file or mic by PJSUA, and Play them on my PC by mplayer => Slow and low pitch. It would be correct if playing it with double rate, saying 16000Hz. 6. VoIP between PJSUA on Davinci and PJSUA on Linux. => Slow and low pitch. So I infer the problem is triggered by pjsip. How can I dig out them in pjsip? FCWu 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: > Hi, > > Not sure what's going on, but I agree it seems to be a sound device > issue, e.g: it may be actually opened in either stereo or 16KHz. I > found this old post that may be related to this issue: > http://mlblog.osdir.com/linux.davinci/2006-10/msg00283.shtml, as it > says "Davinci OSS driver don't work correctly in mono mode and driver > still return two samples per tick". Perhaps you can check this by > running pjsua in stereo mode (with param '--stereo') and see the mic > recording result. Then you can also check pjsua with clock rate 16KHz. > > In case it *is* sound device issue, you can also search/post a > question in the PortAudio forum (PJSIP uses PortAudio on linux > platform). Moreover, you can implement your own sound device wrapper, > please see http://trac.pjsip.org/repos/wiki/External_Sound_Device for > the build config. > > Regards, > nanang > > > On Thu, Oct 9, 2008 at 1:25 PM, Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw> wrote: >> Hi All, >> >> I experienced pjsua on TI Davinci, which configured ./aconfigure >> --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG >> -mcpu=arm926ej-s' --disable-floating-point and defined "#define >> PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. >> When I run pjsua to record sound of 10 second from mic to a file, the >> file would be extended to 20 second long and sound lower pitch than >> original one. (Play it with double rate, it sound fine.) I suppose >> it's because recording device has wrong recording rate. Anybody knows >> how can I adjust it? >> >> p.s. If I recorded a default ring sound from port 02 (Port #02[ >> 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the >> problem come from recording device (?). >> >> The following are the steps I recorded sound from mic. >> ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav >> --clock-rate=8000 --snd-clock-rate=8000 >> ... >>>>> cc 0 1 >> 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open >> devices @8000 Hz >> 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, >> samples per frame=160, tail length=200 ms, latency=96 ms >> 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) >> >> >> >> Best Regards, >> FCWu >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >

Hi, Just an opinion, those tests show that PJSUA on Davinci records more samples than it is supposed to, and also plays more samples than it is supposed to (test #1 - #4 involved only Davinci where record & play are sync'd, thus they are fine, while test #5 & #6 involved other machine). Or it is just an impression caused by inaccurate pjlib data types definitions for Davinci (e.g: 'pj_int16_t' is defined as 'short', while 'short' in Davinci may be 32 bit length instead of 16 bit). So, I think more tests are needed before conclude something here, e.g: - sndtest (in pjsip-apps/src/samples directory), check if the duration printed by sndtest is about twice longer (e.g by default sndtest duration is 10s, the output duration should be around 10s too). - play a correct 8KHz wav file on Davinci's PJSUA, check if it is played faster. - run pjsua in stereo mode 8KHz, do a recording, it will result a stereo 8KHz file, play that recorded file as mono 8KHz. Regards, nanang 2008/10/10 Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw>: > Hi, > > I upload two wav files to explain my problem. > The file recorded by this > way(http://trac.pjsip.org/repos/wiki/audio-how-to-record) could be > found here http://moon.cse.yzu.edu.tw/~s912356/output_people_8000.wav. > It sound slow and low pitch. The correct one should sound like this > http://moon.cse.yzu.edu.tw/~s912356/output_people_16000.wav. > > I made following testings to clarify my problem. > 1. "cat /dev/dsp > file", and then "cat file > /dev/dsp" => Correct > 2. Record and Play audio by portaudio(testing src is > third_party/portaudio/test/patest_record.c) => Correct > 3. PJSUA Looping-back > testing(http://trac.pjsip.org/repos/wiki/audio-check-loopback) => > Correct > 4. Record any audio either from file or mic by PJSUA, and Play them on > Davinci by PJSUA and sample/playfile => Correct > 5. Record any audio either from file or mic by PJSUA, and Play them on > my PC by mplayer => Slow and low pitch. It would be correct if playing > it with double rate, saying 16000Hz. > 6. VoIP between PJSUA on Davinci and PJSUA on Linux. => Slow and low pitch. > > So I infer the problem is triggered by pjsip. How can I dig out them in pjsip? > > > > FCWu > > > > 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: >> Hi, >> >> Not sure what's going on, but I agree it seems to be a sound device >> issue, e.g: it may be actually opened in either stereo or 16KHz. I >> found this old post that may be related to this issue: >> http://mlblog.osdir.com/linux.davinci/2006-10/msg00283.shtml, as it >> says "Davinci OSS driver don't work correctly in mono mode and driver >> still return two samples per tick". Perhaps you can check this by >> running pjsua in stereo mode (with param '--stereo') and see the mic >> recording result. Then you can also check pjsua with clock rate 16KHz. >> >> In case it *is* sound device issue, you can also search/post a >> question in the PortAudio forum (PJSIP uses PortAudio on linux >> platform). Moreover, you can implement your own sound device wrapper, >> please see http://trac.pjsip.org/repos/wiki/External_Sound_Device for >> the build config. >> >> Regards, >> nanang >> >> >> On Thu, Oct 9, 2008 at 1:25 PM, Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw> wrote: >>> Hi All, >>> >>> I experienced pjsua on TI Davinci, which configured ./aconfigure >>> --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG >>> -mcpu=arm926ej-s' --disable-floating-point and defined "#define >>> PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. >>> When I run pjsua to record sound of 10 second from mic to a file, the >>> file would be extended to 20 second long and sound lower pitch than >>> original one. (Play it with double rate, it sound fine.) I suppose >>> it's because recording device has wrong recording rate. Anybody knows >>> how can I adjust it? >>> >>> p.s. If I recorded a default ring sound from port 02 (Port #02[ >>> 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the >>> problem come from recording device (?). >>> >>> The following are the steps I recorded sound from mic. >>> ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav >>> --clock-rate=8000 --snd-clock-rate=8000 >>> ... >>>>>> cc 0 1 >>> 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open >>> devices @8000 Hz >>> 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, >>> samples per frame=160, tail length=200 ms, latency=96 ms >>> 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) >>> >>> >>> >>> Best Regards, >>> FCWu >>> >>> _______________________________________________ >>> Visit our blog: http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >

Thank you, nanaing. The problem can be solved by adding --stereo param. For the people who want to use pjsip on Davinci, you should add following param in your config file for better audio quality. --stereo --ec-tail=0 --add-codec pcmu --dis-codec speex --dis-codec ilbc --dis-codec GSM --dis-codec G722 --clock-rate=8000 --snd-clock-rate=8000 Here is the remaining results of the tests you mentioned. - In Davinci, 'short' is 16 bits - sndtest output the duration is 11s - playing a 8kHz file on Davinci is centain faster than x86 Best Regards, Fan-Cheng Wu 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: > Hi, > > Just an opinion, those tests show that PJSUA on Davinci records more > samples than it is supposed to, and also plays more samples than it is > supposed to (test #1 - #4 involved only Davinci where record & play > are sync'd, thus they are fine, while test #5 & #6 involved other > machine). Or it is just an impression caused by inaccurate pjlib data > types definitions for Davinci (e.g: 'pj_int16_t' is defined as > 'short', while 'short' in Davinci may be 32 bit length instead of 16 > bit). > > So, I think more tests are needed before conclude something here, e.g: > - sndtest (in pjsip-apps/src/samples directory), check if the duration > printed by sndtest is about twice longer (e.g by default sndtest > duration is 10s, the output duration should be around 10s too). > - play a correct 8KHz wav file on Davinci's PJSUA, check if it is played faster. > - run pjsua in stereo mode 8KHz, do a recording, it will result a > stereo 8KHz file, play that recorded file as mono 8KHz. > > Regards, > nanang > > > 2008/10/10 Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw>: >> Hi, >> >> I upload two wav files to explain my problem. >> The file recorded by this >> way(http://trac.pjsip.org/repos/wiki/audio-how-to-record) could be >> found here http://moon.cse.yzu.edu.tw/~s912356/output_people_8000.wav. >> It sound slow and low pitch. The correct one should sound like this >> http://moon.cse.yzu.edu.tw/~s912356/output_people_16000.wav. >> >> I made following testings to clarify my problem. >> 1. "cat /dev/dsp > file", and then "cat file > /dev/dsp" => Correct >> 2. Record and Play audio by portaudio(testing src is >> third_party/portaudio/test/patest_record.c) => Correct >> 3. PJSUA Looping-back >> testing(http://trac.pjsip.org/repos/wiki/audio-check-loopback) => >> Correct >> 4. Record any audio either from file or mic by PJSUA, and Play them on >> Davinci by PJSUA and sample/playfile => Correct >> 5. Record any audio either from file or mic by PJSUA, and Play them on >> my PC by mplayer => Slow and low pitch. It would be correct if playing >> it with double rate, saying 16000Hz. >> 6. VoIP between PJSUA on Davinci and PJSUA on Linux. => Slow and low pitch. >> >> So I infer the problem is triggered by pjsip. How can I dig out them in pjsip? >> >> >> >> FCWu >> >> >> >> 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: >>> Hi, >>> >>> Not sure what's going on, but I agree it seems to be a sound device >>> issue, e.g: it may be actually opened in either stereo or 16KHz. I >>> found this old post that may be related to this issue: >>> http://mlblog.osdir.com/linux.davinci/2006-10/msg00283.shtml, as it >>> says "Davinci OSS driver don't work correctly in mono mode and driver >>> still return two samples per tick". Perhaps you can check this by >>> running pjsua in stereo mode (with param '--stereo') and see the mic >>> recording result. Then you can also check pjsua with clock rate 16KHz. >>> >>> In case it *is* sound device issue, you can also search/post a >>> question in the PortAudio forum (PJSIP uses PortAudio on linux >>> platform). Moreover, you can implement your own sound device wrapper, >>> please see http://trac.pjsip.org/repos/wiki/External_Sound_Device for >>> the build config. >>> >>> Regards, >>> nanang >>> >>> >>> On Thu, Oct 9, 2008 at 1:25 PM, Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw> wrote: >>>> Hi All, >>>> >>>> I experienced pjsua on TI Davinci, which configured ./aconfigure >>>> --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG >>>> -mcpu=arm926ej-s' --disable-floating-point and defined "#define >>>> PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. >>>> When I run pjsua to record sound of 10 second from mic to a file, the >>>> file would be extended to 20 second long and sound lower pitch than >>>> original one. (Play it with double rate, it sound fine.) I suppose >>>> it's because recording device has wrong recording rate. Anybody knows >>>> how can I adjust it? >>>> >>>> p.s. If I recorded a default ring sound from port 02 (Port #02[ >>>> 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the >>>> problem come from recording device (?). >>>> >>>> The following are the steps I recorded sound from mic. >>>> ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav >>>> --clock-rate=8000 --snd-clock-rate=8000 >>>> ... >>>>>>> cc 0 1 >>>> 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open >>>> devices @8000 Hz >>>> 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, >>>> samples per frame=160, tail length=200 ms, latency=96 ms >>>> 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) >>>> >>>> >>>> >>>> Best Regards, >>>> FCWu >>>> >>>> _______________________________________________ >>>> Visit our blog: http://blog.pjsip.org >>>> >>>> pjsip mailing list >>>> pjsip@lists.pjsip.org >>>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>>> >>> >>> _______________________________________________ >>> Visit our blog: http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >>> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >

Hi, So it seems the sound device in Davinci is actually always opened in stereo mode, even when --stereo is not specified (not sure whether the problem starts in sound driver or PA though). Regards, nanang 2008/10/11 Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw>: > Thank you, nanaing. The problem can be solved by adding --stereo param. > > For the people who want to use pjsip on Davinci, you should add > following param in your config file for better audio quality. > --stereo > --ec-tail=0 > --add-codec pcmu > --dis-codec speex > --dis-codec ilbc > --dis-codec GSM > --dis-codec G722 > --clock-rate=8000 > --snd-clock-rate=8000 > > Here is the remaining results of the tests you mentioned. > - In Davinci, 'short' is 16 bits > - sndtest output the duration is 11s > - playing a 8kHz file on Davinci is centain faster than x86 > > Best Regards, > Fan-Cheng Wu > > > > 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: >> Hi, >> >> Just an opinion, those tests show that PJSUA on Davinci records more >> samples than it is supposed to, and also plays more samples than it is >> supposed to (test #1 - #4 involved only Davinci where record & play >> are sync'd, thus they are fine, while test #5 & #6 involved other >> machine). Or it is just an impression caused by inaccurate pjlib data >> types definitions for Davinci (e.g: 'pj_int16_t' is defined as >> 'short', while 'short' in Davinci may be 32 bit length instead of 16 >> bit). >> >> So, I think more tests are needed before conclude something here, e.g: >> - sndtest (in pjsip-apps/src/samples directory), check if the duration >> printed by sndtest is about twice longer (e.g by default sndtest >> duration is 10s, the output duration should be around 10s too). >> - play a correct 8KHz wav file on Davinci's PJSUA, check if it is played faster. >> - run pjsua in stereo mode 8KHz, do a recording, it will result a >> stereo 8KHz file, play that recorded file as mono 8KHz. >> >> Regards, >> nanang >> >> >> 2008/10/10 Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw>: >>> Hi, >>> >>> I upload two wav files to explain my problem. >>> The file recorded by this >>> way(http://trac.pjsip.org/repos/wiki/audio-how-to-record) could be >>> found here http://moon.cse.yzu.edu.tw/~s912356/output_people_8000.wav. >>> It sound slow and low pitch. The correct one should sound like this >>> http://moon.cse.yzu.edu.tw/~s912356/output_people_16000.wav. >>> >>> I made following testings to clarify my problem. >>> 1. "cat /dev/dsp > file", and then "cat file > /dev/dsp" => Correct >>> 2. Record and Play audio by portaudio(testing src is >>> third_party/portaudio/test/patest_record.c) => Correct >>> 3. PJSUA Looping-back >>> testing(http://trac.pjsip.org/repos/wiki/audio-check-loopback) => >>> Correct >>> 4. Record any audio either from file or mic by PJSUA, and Play them on >>> Davinci by PJSUA and sample/playfile => Correct >>> 5. Record any audio either from file or mic by PJSUA, and Play them on >>> my PC by mplayer => Slow and low pitch. It would be correct if playing >>> it with double rate, saying 16000Hz. >>> 6. VoIP between PJSUA on Davinci and PJSUA on Linux. => Slow and low pitch. >>> >>> So I infer the problem is triggered by pjsip. How can I dig out them in pjsip? >>> >>> >>> >>> FCWu >>> >>> >>> >>> 2008/10/10 Nanang Izzuddin <nanang@pjsip.org>: >>>> Hi, >>>> >>>> Not sure what's going on, but I agree it seems to be a sound device >>>> issue, e.g: it may be actually opened in either stereo or 16KHz. I >>>> found this old post that may be related to this issue: >>>> http://mlblog.osdir.com/linux.davinci/2006-10/msg00283.shtml, as it >>>> says "Davinci OSS driver don't work correctly in mono mode and driver >>>> still return two samples per tick". Perhaps you can check this by >>>> running pjsua in stereo mode (with param '--stereo') and see the mic >>>> recording result. Then you can also check pjsua with clock rate 16KHz. >>>> >>>> In case it *is* sound device issue, you can also search/post a >>>> question in the PortAudio forum (PJSIP uses PortAudio on linux >>>> platform). Moreover, you can implement your own sound device wrapper, >>>> please see http://trac.pjsip.org/repos/wiki/External_Sound_Device for >>>> the build config. >>>> >>>> Regards, >>>> nanang >>>> >>>> >>>> On Thu, Oct 9, 2008 at 1:25 PM, Fan-Cheng Wu �����\ <fcwu.cs95g@nctu.edu.tw> wrote: >>>>> Hi All, >>>>> >>>>> I experienced pjsua on TI Davinci, which configured ./aconfigure >>>>> --host=armv5tl-montavista-linuxeabi CFLAGS='-O2 -msoft-float -DNDEBUG >>>>> -mcpu=arm926ej-s' --disable-floating-point and defined "#define >>>>> PJSUA_DEFAULT_CLOCK_RATE (8000)" on config_site.h. >>>>> When I run pjsua to record sound of 10 second from mic to a file, the >>>>> file would be extended to 20 second long and sound lower pitch than >>>>> original one. (Play it with double rate, it sound fine.) I suppose >>>>> it's because recording device has wrong recording rate. Anybody knows >>>>> how can I adjust it? >>>>> >>>>> p.s. If I recorded a default ring sound from port 02 (Port #02[ >>>>> 8KHz/20ms/1] ring transmitting to: ), it's also correct. So the >>>>> problem come from recording device (?). >>>>> >>>>> The following are the steps I recorded sound from mic. >>>>> ./pjsua-armv5tl-montavista-linux-gnueabi --rec-file=output.wav >>>>> --clock-rate=8000 --snd-clock-rate=8000 >>>>> ... >>>>>>>> cc 0 1 >>>>> 07:49:01.566 pjsua_media.c pjsua_set_snd_dev(): attempting to open >>>>> devices @8000 Hz >>>>> 07:49:01.758 ec0x150df8 AEC created, clock_rate=8000, channel=1, >>>>> samples per frame=160, tail length=200 ms, latency=96 ms >>>>> 07:49:01.759 conference.c Port 3 (ring) transmitting to port 1 (output.wav) >>>>> >>>>> >>>>> >>>>> Best Regards, >>>>> FCWu >>>>> >>>>> _______________________________________________ >>>>> Visit our blog: http://blog.pjsip.org >>>>> >>>>> pjsip mailing list >>>>> pjsip@lists.pjsip.org >>>>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>>>> >>>> >>>> _______________________________________________ >>>> Visit our blog: http://blog.pjsip.org >>>> >>>> pjsip mailing list >>>> pjsip@lists.pjsip.org >>>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>>> >>>> >>> >>> _______________________________________________ >>> Visit our blog: http://blog.pjsip.org >>> >>> pjsip mailing list >>> pjsip@lists.pjsip.org >>> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >>> >>> >> >> _______________________________________________ >> Visit our blog: http://blog.pjsip.org >> >> pjsip mailing list >> pjsip@lists.pjsip.org >> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org >> >> > > _______________________________________________ > Visit our blog: http://blog.pjsip.org > > pjsip mailing list > pjsip@lists.pjsip.org > http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org > >