> Simon,
>
> what it takes speex to run on ARM.
>
> Thanks and Regards,
> HarishKumar.V
>
> On Fri, Jul 3, 2009 at 12:45 PM, Harishkumar V
<
address@hidden>wrote:
> > Simon,
> >
> > Using PCMU, PCMA and ILBC, it works, with the original code, no changes
> > in alsa period and alsa period size.
> >
> > when i am using speex, it fails with the following message,
> >
> > ortp-error-snd_pcm_avail_update: Broken pipe
> > ortp-error-*** alsa_can_read fixup, trying to recover
> > ALSA lib pcm.c:7232:(snd_pcm_recover) overrun occured
> > ortp-error-snd_pcm_avail_update: Broken pipe
> > ortp-error-*** alsa_can_read fixup, trying to recover
> > ALSA lib pcm.c:7232:(snd_pcm_recover) overrun occured
> >
> >
> > whats the difference between speex and other 3 codecs. In lan, the
> > quality is good.
> > How the quality will be over the internet when using above 3 codecs.
> >
> > Is it makes any difference using the above 3 codecs without speex.
> >
> > Thanks and Regards,
> > HarishKumar.V
> >
> > On Fri, Jul 3, 2009 at 12:04 PM, Code Warrior <
address@hidden> wrote:
> >> CPU utilization is under 3% for Linphone. So that really doesn't seem to
> >> be the issue.
> >>
> >> On Fri, Jul 3, 2009 at 2:19 AM, Simon Morlat
<
address@hidden>wrote:
> >>> Hi,
> >>>
> >>> Did you try by uncomenting the #define EPIPE_BUGFIX ?
> >>> In order to configure codecs, simply edit ~/.linphonerc and set
> >>> 'enable=0' in
> >>> each codec section (named audio_codec_%i). Only keep pcmu and pcma,
> >>> because
> >>> they consume very few cpu.
> >>> My feeling is that when in call, the cpu consumption causes an overrun
> >>> or underrun at the alsa driver level, which cannot be recovered as it
> >>> should with
> >>> other cards.
> >>> When aplay/arecord or mediastream, the cpu consumption is lesser and
> >>> the problem does not appear.
> >>>
> >>> Simon
> >>>
> >>> Le jeudi 2 juillet 2009 17:15:05, Harishkumar V a écrit :
> >>> > Simon,
> >>> >
> >>> > I debugged bit more, i found this one,
> >>> >
> >>> > snd_pcm_avail_update in alsa_can_read function, where it returns
> >>> > EPIPE error and snd_pcm_recover also fails during having a call
> >>> > between me
> >>>
> >>> and
> >>>
> >>> > other over the lan network. but the same snd_pcm_avail_update returns
> >>>
> >>> the
> >>>
> >>> > available count of samples if i use mediastream as mentioned in
> >>>
> >>> previous
> >>>
> >>> > mail and works fine.
> >>> >
> >>> > I will let u know the codec being used shortly tomorrow, as i left
> >>> > for
> >>>
> >>> the
> >>>
> >>> > home.
> >>> >
> >>> > What codec i should use for the better performance for the both A/V.
> >>>
> >>> how to
> >>>
> >>> > configure it.
> >>> >
> >>> > Thanks and Regards,
> >>> > HarishKumar.V
> >>> >
> >>> > On Thu, Jul 2, 2009 at 7:06 PM, Simon Morlat
> >>>
> >>> <
address@hidden>wrote:
> >>> > > Le jeudi 2 juillet 2009 12:54:33, Harishkumar V a écrit :
> >>> > > > Hi Simon,
> >>> > > >
> >>> > > > when i run "mediastream --local 8000 --remote
> >>> > > > 127.0.0.1:8000--payload
> >>> > >
> >>> > > 0"
> >>> > >
> >>> > > > locally with the default alsa periods and alsa period size. its
> >>>
> >>> fine,
> >>>
> >>> > > > no probs. i can hear voice recorded being playback.
> >>> > > >
> >>> > > > When i try to call or somebody calls me, then after answering the
> >>>
> >>> call,
> >>>
> >>> > > > this error happens, and the voice comes from my ARM platform to
> >>> > > > PC,
> >>>
> >>> but
> >>>
> >>> > > so
> >>> > >
> >>> > > > many packet loss, if i speak 10 words, only 2 words arrive at the
> >>>
> >>> other
> >>>
> >>> > > end
> >>> > >
> >>> > > > after delay.
> >>> > >
> >>> > > Seems you are running out of cpu. What codec is used ?
> >>> > >
> >>> > > Simon
> >>> > >
> >>> > > > Thanks and Regards,
> >>> > > > HarishKumar.V
> >>> > > >
> >>> > > > On Thu, Jul 2, 2009 at 3:46 PM, Harishkumar V
> >>> > >
> >>> > > <
address@hidden>wrote:
> >>> > > > > Hi Simon,
> >>> > > > >
> >>> > > > > i can able to understand period_size. In what order does
> >>>
> >>> ALSA_PERIODS
> >>>
> >>> > > is
> >>> > >
> >>> > > > > to be incremented, how is co-related with the ALSA_PERIOD_SIZE.
> >>> > > > >
> >>> > > > > Thanks and Regards,
> >>> > > > > HarishKumar.V
> >>> > > > >
> >>> > > > > On Thu, Jul 2, 2009 at 1:43 PM, Simon Morlat
> >>> > >
> >>> > > <
address@hidden>wrote:
> >>> > > > >> Ok, big problems with the alsa driver.
> >>> > > > >> You can relax the stress on the driver by requesting larger
> >>>
> >>> buffer:
> >>> > > > >> see the comments at the head of mediastreamer2/src/alsa.c:
> >>> > > > >>
> >>> > > > >> /*in case of troubles with a particular driver, try
> >>> > > > >> incrementing ALSA_PERIOD_SIZE
> >>> > > > >> to 512, 1024, 2048, 4096...
> >>> > > > >> then try incrementing the number of periods*/
> >>> > > > >> #define ALSA_PERIODS 8
> >>> > > > >> #define ALSA_PERIOD_SIZE 256
> >>> > > > >>
> >>> > > > >> /*uncomment the following line if you have problems with an
> >>> > > > >> alsa
> >>> > >
> >>> > > driver
> >>> > >
> >>> > > > >> having sound quality trouble:*/
> >>> > > > >> /*#define EPIPE_BUGFIX 1*/
> >>> > > > >>
> >>> > > > >> First try setting ALSA_PERIOD_SIZE to 512 then 1024, then
> >>> > > > >> 2048,
> >>>
> >>> try
> >>>
> >>> > > > >> incrementing ALSA_PERIODS, then uncomment the #define
> >>>
> >>> EPIPE_BUGFIX .
> >>>
> >>> > > > >> Incrementing periods and periodsize also increments the
> >>> > > > >> latency
> >>>
> >>> of
> >>>
> >>> > > > >> the system,
> >>> > > > >> ie the delay for voice to be captured and playback.
> >>> > > > >>
> >>> > > > >> Simon
> >>> > > > >>
> >>> > > > >> Le jeudi 2 juillet 2009 09:46:36, Harishkumar V a écrit :
> >>> > > > >> > hi,
> >>> > > > >> >
> >>> > > > >> > I am running linphone-3.1.0 on my ARM platform. i am facing
> >>>
> >>> this
> >>>
> >>> > > > >> problem's,
> >>> > > > >>
> >>> > > > >> > when i execute linphonec.
> >>> > > > >> >
> >>> > > > >> > #linphonec
> >>> > > > >> >
> >>> > > > >> > Ready
> >>> > > > >> > Warning: video is disabled in linphonec, use -V or -C or -D
> >>> > > > >> > to
> >>> > >
> >>> > > enable.
> >>> > >
> >>> > > > >> > linphonec> <
sip:address@hidden <
address@hidden> <
> >>>
> >>>
address@hidden <
address@hidden>> <
> >>>
> >>> > >
address@hidden <
address@hidden> <
> >>>
> >>>
address@hidden <
address@hidden>>> <
> >>>
> >>> > > > >>
address@hidden <
address@hidden> <
> >>>
> >>>
address@hidden <
address@hidden>> <
> >>>
> >>> > >
address@hidden <
address@hidden> <
> >>>