Hi, news from my bug :
I found that when I play with ALSA_PERIODS and ALSA_PERIOD_SIZE
in alsa.c it changes the output of my sound. The sound is still
bad but I clearly see that it changes the frequency of the
stutter/distortion.
So my bug is clearly situated in mediastreamer. When I launch
linphone (verbose) I can see that log :
Strange, sound card aic31xx-audio does not seems to be capable of
anything, retrying with plughw...
And in the code there is a comment saying that mediastreamer trys
to use plughw to workaround an alsa bug. So my question is simple
: What can I do with my alsa driver to avoid this comment ?
What do you think of OSS ?
Thanks !
Le 01/02/2018 à 12:10, Thomas Dalmayrac
a écrit :
Hey !
I discussed about my trouble on chat.freenode.net #linphone
yesturday and I had many ideas from there. One of the ideas is
to revert the mediastreamer library, what do you think of that ?
Thomas.
Le 30/01/2018 à 15:55, Thomas
Dalmayrac a écrit :
Hi everybody !
I'm working on a on an i.MX6 (freescale) based product under
a buildroot personnalized linux, and I try to use linphonec
with it, but I've got troubles when I'm in call (voice is
hatched). See the record I've made from the call : https://vocaroo.com/i/s1RrgXBSXIaT.
The wireshark : https://www.cloudshark.org/captures/9fd82b8b4e60
To test and isolate the issue of hatched voice I played an
example sound from an asterisk server - I correctly establish
a call but every ~35sec the voice is hatched for 5sec. The
codec used for the call is PCMU, and the soundcard is an I2S
with a codec TI TLV320AIC3110 (http://www.ti.com/product/TLV320AIC3110).
- The sound is correctly
received by the softphone. We put linphone in file
mode (without sound card - "soundcard use files"
"record") to record the audio received by the phone
before it is sent to the lower layers. It is about the 5th
second that we hear a fault on the loudspeakers
(like the sound given in intro see https://vocaroo.com/i/s1RrgXBSXIaT),
while on the recorded sample, we do not hear any
defect .
- The disturbance
occurs every 35 seconds or so.
- The disturbance is
insensitive to the CPU load (tested with stress), or the
network load (tested with iperf)
- The sound is correctly
transported by the network, there are no losses and if
we sniff the packets received by the handset we can play
the sound without worry via wireshark (https://www.cloudshark.org/captures/9fd82b8b4e60)
- The codecs are correctly
managed (we will use the PCMU G711-Ulaw, tested G722 and
PCMA ok)
- The
same example sound is correctly decoded / played
by aplay if we download it.
Perhaps an important fact : Everytime I start linphonec I got
these messages :
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 0
ALSA lib control.c:954:(snd_ctl_open_noupdate) Invalid CTL
default:0
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 0
ALSA lib pcm.c:2266:(snd_pcm_open_noupdate) Unknown PCM
default:0
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 0
ALSA lib pcm.c:2266:(snd_pcm_open_noupdate) Unknown PCM
default:0
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 1
ALSA lib control.c:954:(snd_ctl_open_noupdate) Invalid CTL
default:1
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 1
ALSA lib pcm.c:2266:(snd_pcm_open_noupdate) Unknown PCM
default:1
ALSA lib conf.c:4738:(snd_config_expand) Unknown parameters 1
ALSA lib pcm.c:2266:(snd_pcm_open_noupdate) Unknown PCM
default:1
We think
that the concern comes from the lower layers used by
the linphone. Can there be incompatibilities
with MediaStreamer2? Does all
this bring up something that could help us solve our
problem?
Thank you for every single idea you'll give to us :)
--
Thomas Dalmayrac
_______________________________________________
Linphone-developers mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/linphone-developers
--
Thomas Dalmayrac
_______________________________________________
Linphone-developers mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/linphone-developers
--
Thomas Dalmayrac