Closed GoogleCodeExporter closed 9 years ago
[deleted comment]
Please provide any additional information below.
Update:
I am using CSipSimple-r1038-tls on all the mobiles now. My first idea with the
older version was a mistake (I tried with r1025 and update then to the newest).
The OS is still the same.
Calls:
Samsung Galaxy S2 calling HTC Wildfire -> Sound ok (most of the time, not with
all testcalls)
the other way round: HTC Wildfire calling Samsung Galaxy S2 -> Sound distorted
(I tried both mobiles)
HTC Wildfire calling HTC Wildfire -> Sound distorted (both mobiles)
all cyanogenmod settings should show standard, all mobiles in the same wifi.
do you need any logs for solving this issue? if yes, which mobile and which
testcase :)
thank you
Original comment by automate...@gmail.com
on 11 Oct 2011 at 7:56
1.) I update with CSipSimple-r1040-tls
2.) I did six testcases; only one is ok. all the others have an issue. here are
the details:
A) SGS2 calling HTC Wildfire: call ok
B) HTC Wildfire calling SGS2: voice out of the HTC loudspeaker is ok, the one
coming out of the SGS2 is distorted
C) SGS2 calling 2nd HTC Wildfire: voice out of the SGS2 loudspeaker is ok, the
one coming out of the HTC is distorted
D) 2nd HTC Wildfire calling SGS2: voice out of the HTC loudspeakr is ok, the
one coming out of the SGS2 is distorted
E)HTCW1 calling HTCW2: The called one receives the distorted voice. The caller
mobile's sound is good.
F)HTCW2 calling HTCW1 -> same like above: The called one receives the distorted
voice signal as well. The voice coming out of the caller's loudspeaker is fine.
I will start creating logs and posting it as a follow up
Original comment by automate...@gmail.com
on 12 Oct 2011 at 7:43
CSipSimple Error-Log reports of all three
Original comment by automate...@gmail.com
on 12 Oct 2011 at 8:13
Attachments:
This is how the audio is - see attached.
It is the sound of a hair dryer. The other side can hear it clearly.
Meanwhile i've updated the HTC Wilfire to cyanogenmod-7.1.0.1 and
CSipSimple-r1047-tls now :).
Original comment by automate...@gmail.com
on 16 Oct 2011 at 1:18
Attachments:
sorry, here the attachment :)
Original comment by automate...@gmail.com
on 16 Oct 2011 at 1:19
Attachments:
Do you get same results if you disable all codecs but PCMU/PCMA.
G722 use a lot of CPU and maybe resampling + high clock rate lock Wilfire's CPU.
If only PCMU/PCMA helps, it could be interesting to test other things on the
wilfire. I'll tell you. If g722 is not suitable for Wilfire (and I guess all
armv6 devices without floating point cpu support, I will probably consider
disable it by default on this device).
The fact that it is due to codec, could explain why you don't get same results
for incoming/outgoing calls cause codec negociation can result different codec.
Original comment by r3gis...@gmail.com
on 16 Oct 2011 at 3:11
Thank you mate, good idea. After picking a different codec it is back to
normal. Deactivating it in the standard config will be a good solution to this
imo.
It seems g722 is too much for the Wildfire. Speex or PCMU/PCMA is ok though.
Allow me a comment on the side:
Even if I deacitvate all the codecs - for narrow and wideband - except speex32,
it will pick speex16. is it deliverately - kind of a shrink down to the next
possible one?
Original comment by automate...@gmail.com
on 16 Oct 2011 at 7:49
Ok, for g722. I'll disable that by default on cpu with low processing
capabilities.
About speex, weird, I'll check on my side. It is possible that it select
speex16 if another speex is enabled and that the other side prefer speex16. But
if both devices has only speex32, it should use speex32.
I'll check that, thanks for pointing this out.
Original comment by r3gis...@gmail.com
on 16 Oct 2011 at 8:09
This issue was closed by revision r1051.
Original comment by r3gis...@gmail.com
on 16 Oct 2011 at 8:17
Actually it is still broken.
I've tried r1058 with GalaxyS2 and Nexus One.
The sound is almost always distorted on a phone I am calling to.
Nick
Original comment by nickusti...@gmail.com
on 20 Oct 2011 at 5:27
Original issue reported on code.google.com by
automate...@gmail.com
on 7 Oct 2011 at 5:32