Closed GoogleCodeExporter closed 9 years ago
the ringing was deacitvated. Now I find out, after opening the settings dialog
but change noting, it was working. Now I will check out, if it happens after I
switched from WLAN to UMTS and back. Outgoing is working, incoming has no
sound. With ringing and vibrate the same.
Original comment by m...@ufritz.de
on 5 Sep 2010 at 1:39
Ok. I found the behaviour:
After returning from UMTS to wlan, csipsimple re register the account (can see
it on the symbol). But incoming calls have no sound (earspeaker).
If I do deactivate and then reactivate the account (account settings). There is
still no sound.
If I open the settings dialog, go to settings -> media and then go back, not
change anything, there is a re register again, but then I will have sound.
Maybe need to uninstall complete and install again?
Original comment by m...@ufritz.de
on 5 Sep 2010 at 2:01
Ok !
That's issue 126 !
STUN is not re-resolved and you probably re-register with the old ip address.
(Going to settings and then back, restart all the stack while ip changes
detection only re-register - and that's not enough if public IP has changed).
Original comment by r3gis...@gmail.com
on 5 Sep 2010 at 2:43
ok. can confirm this, after disabling STUN and enabling ICE (settings -> media)
it is working. So this was the STUN Server setting.
Original comment by m...@ufritz.de
on 5 Sep 2010 at 3:09
I'll try to fix it in next release. I'll follow advises from pjsip (the native
sip stack on which csipsimple is based) and fully restart all the sip stack on
network change.
Will probably fix the issue in a first time but will consume more CPU (and so
battery) when network change is detected (but I assume there is not a lot of
network change so that can be considered as a minor impact).
Original comment by r3gis...@gmail.com
on 5 Sep 2010 at 3:25
> and fully restart all the sip stack on network change
maybe the fully restart is only made, if STUN is activated?
In my case. I have a lot of network changes, because if I go outside (my
garden) and go inside back, I have already 2 changes, and I go outside a lot at
day.
Original comment by m...@ufritz.de
on 5 Sep 2010 at 3:58
I don't remember to tell that with older version and with newest TCP transport
doesn't work. If I unselect UDP and only TCP is selected will csipsimple crash
or it is hard to get back to settings menu (because it want to jump to androids
default screen).
Original comment by jar...@gmail.com
on 6 Sep 2010 at 2:29
To galaxy S owners :
I've released a -12-29 version.
It add an option to change a parameter called ptime. I would like to test and
change it to 20.
I ask this because, -12-27 that solve the micro issue for you has both a
changed ptime value (this value is linked to latency) (from 20 to 30), and also
the implementation tweak that wait before asking micro frames. If you confirm
that 20 works as fine as 30 I'll put 20 as the default value (latency will be
better - I mean audio latency, small compared to network latency -> so you'll
not ear a big difference).
@jarsiv : for tcp/udp issue can you send me your logs?
Well in all case both TCP and UDP ~should~ be always activated (unless it's a
really specific configuration), if you want to use TCP you should change it
into the account configuration, but still interested in crash logs :) unless
it's an ANR (application not responding) and so I think that I know the reason.
(The application doesn't thread everything yet and sometimes it tries to
register/resolve things in the user interface thread which mean a lack of
responsiveness. So if your STUN server/registrar server or another server
doesn't support TCP it can wait for a network timeout in the UI thread. This is
a big point I have to fix but I keep it in mind.
@mail : New version restart everything. As you said, it's not really optimized.
I don't know exactly what impact it could have on the battery. I don't think
that it's so huge but for now it will solve *a lot* of problems (included wifi
lock that could remains - and maybe drain more battery). Let me know if you
observe a big difference with this version. Anyway I'll need to optimize this
later but could help me on the priority of the task.
Original comment by r3gis...@gmail.com
on 6 Sep 2010 at 8:23
I tested ptime 20 and I don't hear any differens. So I think it's ok set it
default to 20.
Also mailed logs of tcp issue.
Original comment by jar...@gmail.com
on 7 Sep 2010 at 6:55
[deleted comment]
I found a new sound issue:
If i receive a call, the speaker-phone on the back is on (hands free speaking).
If I make a call, it is off (correct).
I can not switch it off or on.
This also produce an echo.
Original comment by m...@ufritz.de
on 7 Sep 2010 at 8:37
Is it possible to add feature that while you are calling and phone is on your
ear csipsimple will shut down the display. And after you tilt the phone to stop
calling it will put on display?
Now display is on whole time so it will drain battery quite fast.
Original comment by jar...@gmail.com
on 9 Sep 2010 at 6:10
comment 62 seems to be in the wrong place.
But anyway, the display on the galaxy just dims, if the sensor comes close to
the head.
With the android standard dialer (GSM), the display dims more, goes to black
dark.
With sipdroid, the display shut down after timeoff, after the call you have to
unlock the display an can hang up
How to tilt the phone to stop calling?
Original comment by m...@ufritz.de
on 9 Sep 2010 at 8:19
Just try to untick the option "Keep awake while in call" in UI settings. It's a
work around for nexus one.
In future release for other devices than nexus one it will be un-ticked but for
now since it has a deep impact on communication on nexus one it is by default
enabled.
Besides :
- I think that the tilt to stop should be done by a third party app ;)
- I'll try to implement soon the real proximity sensor usage (screen goes off
when proximity sensor is activated) - But it requires the usage of a private
android API. So for now, *if you disable the "keep awake while in call" option*
=> Screen will goes off when near of your head after a while (same time than
the one configured in standard case). In the future (with private api) it will
goes off immediately (if it is not the nexus one on wifi !)
Original comment by r3gis...@gmail.com
on 9 Sep 2010 at 9:20
Hi There r3gis.3R,
first I want to say that I love your app, it integrates flawlessly with my
Android. I am having a problem though, when I make a call there is no sound on
my ear piece, or headphones, mic works great, the person on the other line can
hear me no problem but I can not hear the callee, I have change the codecs, I
have follow all and any instruction and nothing resolve the issue. i would
really like some advise, my phone is HTC Desire (nortamerican model), runing
android 2.1 sip provider freephoneline.ca (I have used other sip clients
without problem with this sip provider). Thank you.
Original comment by enciso.d...@gmail.com
on 10 Sep 2010 at 1:09
@enciso : really strange since should works fine with HTC. Probably a
configuration issue. Can you provide me some logs (follow the wiki page about
how to collect logs). I'll open a new bug/issue for your one when I'll get the
logs.
@galaxyS owners : I modified the way audio is routed to speaker / earpiece in
-12-30, can probably help to solve the issue. -You should probably reboot phone
after installing the new version in order to ensure audio API is in a good
state when you'll do tries with new version-
@X10 and i7500 owners : does last version helps with previous crashes? I guess
that what I did to make it work on Galaxy S (and alsa driver) would help to
avoid audio stack crash on X10 & i7500.
Original comment by r3gis...@gmail.com
on 10 Sep 2010 at 9:17
@enciso : was not able to reply to your mail (I get a 550 550 No such user
delivery failure mail). But can you re do the test with logs activated (see
http://code.google.com/p/csipsimple/wiki/HowToCollectLogs the second paragraph
"How to turn on logging"). Will help me to know what is happening :).
Original comment by r3gis...@gmail.com
on 10 Sep 2010 at 10:09
@enciso : thanks we will continue on issue 209 > there is some extra question
for you ;)
Original comment by r3gis...@gmail.com
on 10 Sep 2010 at 10:21
First short test on the galaxy.
The speaker on the back is now off by default. Great!
Switching between the ear and back speaker is now the same behaviour like on
sipdroid. To switch on is working, to switch off again, does not work. But this
is not really important.
At the first test I had no sound again, only after opening the settings screen
and close it again. STUN is activated, but this issue is fixed already. I will
keep an eye on it, maybe I will find a way to reproduce it.
Original comment by m...@ufritz.de
on 10 Sep 2010 at 10:48
Great, thanks a lot for the test.
For the switch off again issue there is something I can try to solve the issue.
It sounds like the one I get on bluetooth and simply stop and restart the
streaming with good params set before opening the stream made things work.(
since for now it is done while streaming is done and switching from earpiece to
speaker doesn't stop audio stream)
The same technique could maybe solve the speaker switching issue.
For reference, still issues on X10 (thanks Mark for the report).
I'll try to test on a colleague X10 next Monday (at work colleagues start to
own android phones....samsung S (one received it today), a X10, and a htc
desire :)... so will be easier for me to debug issues on these device... just
have to convince to test my app on their phone for a couple of hour ;) )
Original comment by r3gis...@gmail.com
on 10 Sep 2010 at 11:09
I have the same problem with Comment 61. When someone call to me, their sound
is always on the back speaker. And when I make a call, the sound is ear
speaker(that is correct). I cannot mute it also. I use samsung galaxy s in
Thailand. I try both with CSipSimple_0.00-12-29.apk and
CSipSimple_0.00-12-30.apk. I also reboot the phone after installed.
ps. I love your app so much, I will wait it to work correctly with my galaxy s.
:)
Original comment by authapon
on 11 Sep 2010 at 10:59
@authapon : are you sure version 0.00-12-30 doesn't solve the issue? (At least
for the default behavior, switching is still broken, but I'm working on it).
For the mic mute, it is quietly sure that it work properly now with 0.00-12-30
: I changed the method and now it's not managed by the android system but by
the sip stack which is more reliable and doesn't depend on the device.
So I guess something went wrong when you installed the 0.00-12-30 you had the
micro mute issue. Try to fully uninstall the app and reinstall the 0.00-12-30
and let me know if you are not in the case of comment 69 now.
Original comment by r3gis...@gmail.com
on 11 Sep 2010 at 2:42
@r3gis.3R : I try to reinstall the version 0.00-12-30 again and reboot my
phone. It still doesn't solve my problem. My back speaker is on when someone
call to me and the mute function is not working too. I am not sure what wrong
with my phone or my sip server. T_T
I try with fring and it is working correctly. I saw that fring will enter to
silent mode(silent mode notification is shown in the task bar) when I accept
the calling. I think that may be the secret for my phone.
Original comment by authapon
on 11 Sep 2010 at 5:44
@r3gis.3R : Oh..I have some misunderstand, the mute function is working
correctly. Now I have only the back speaker problem. :)
Original comment by authapon
on 11 Sep 2010 at 5:58
Ok. Yesterday I made a short test and the back speaker was off. But I spoke not
very loud, because I didn't wake up my kids, but I was sure all is fine.
I don't know what was happens yesterday night...maybe I was sleeping already.
But it is true, the backspeaker is on all time :-(
Tested today again.
The Mute Button is working.
Original comment by m...@ufritz.de
on 11 Sep 2010 at 8:11
@mail : maybe you where in ringing mode silent yesterday?
As authapon proposed, I'll try to set ringer mode to silent during the call
(really good idea, could be consistent with observed behavior).
So can be interesting to test that.
I'll also implement the workaround for switching between speaker and earpiece.
Original comment by r3gis...@gmail.com
on 11 Sep 2010 at 8:19
Yes, you are right, just tested it.
If the ringer is silent, the sound comes over the ear speaker.
If the ringer is not silent, the sound comes over the back speaker.
Original comment by m...@ufritz.de
on 11 Sep 2010 at 8:39
Well, 0.00-12-31
It set ringer mode to silent while in call. Let me know if it helps to solve
the speakerphone issue.
I've also changed the way speakerphone mode is switched (same workaround than
the one for Bluetooth).
Original comment by r3gis...@gmail.com
on 11 Sep 2010 at 8:41
No, it does not help. It seems it has to be silent before it is ringing.
Original comment by m...@ufritz.de
on 11 Sep 2010 at 9:26
Mmhh... There is probably something with ringer. I guess there is some micro
seconds while there is still ringing. I pushed -12-31-bis that tries to
synchronize with the end of the ringer.
Did you test speaker/earpiece switch with -31? (should be the same that 31-bis
on this point)
Original comment by r3gis...@gmail.com
on 11 Sep 2010 at 9:43
Interesting, there is a new behaviour.
If I answer a call, no sound at all is to hear. After pause an resume there is
sound, but again on the back speaker.
Speaker/earpiece switch is not working. But the switch is doing anything,
because if I switch, the back speaker makes a "knack" and it is a little bit
lower noise. This happens with siproid also. So I guess there is a wrong
implementation from Samsung and I guess, this will go, if they bring the
Android 2.2 Update.
The problem is the running back speaker.
Original comment by m...@ufritz.de
on 11 Sep 2010 at 10:14
Huh, amazing... The only thing was to ensure ringer was stopped before starting
audio.
Well, I pushed -12-32. I'm exploring if it's not due to the thread from which
audio mode is set.
If possible could you test and send me logs you get using this one. I don't
think that it will solve the issue, but behavior it will have can help me to
understand what is happening. So if you can test both incoming and outgoing and
both with ringer and without (for incoming).
Thx in advance
Original comment by r3gis...@gmail.com
on 12 Sep 2010 at 4:32
[deleted comment]
sorry mistake, test coming soon
Original comment by m...@ufritz.de
on 12 Sep 2010 at 6:18
ok again:
1. incoming call, ringer off:
no sound, open settings close it and test again.
2. incoming call, ringer off:
all fine
3. incoming call ringer on:
backspeaker on
4. outgoing call:
all is fine
I mail you the logs
Original comment by m...@ufritz.de
on 12 Sep 2010 at 6:46
There are any news for this issue?
Seems to be a huge problem for the alsa implementation of the android phones.
Original comment by m...@ufritz.de
on 17 Sep 2010 at 6:10
I have read that Android Froya 2.2 will release for Galaxy S on 23.9. So
perhaps we should test it first if there is some improve with audio driver.
Original comment by jar...@gmail.com
on 17 Sep 2010 at 6:25
As that release might get delayed.. did anyone test it with the JPH release
floating around lately?
Besides, the sipdroid folks have a similar issue:
http://code.google.com/p/sipdroid/issues/detail?id=341 .
Original comment by p.dreimann@gmail.com
on 17 Sep 2010 at 6:42
I tested with the phone of my colleague for a couple of hours. My conclusion
was there is something really strange in the way they implement speaker on/off
feature.
I tried with breakpoints and even when I wait after the end of the ringing, it
still goes on speaker. (Actually, it goes through earpiece + speaker). In fact,
once speaker is turned on (for ringing for example), it seems to be really
difficult to turn it back to off then. Android API should allow that
flawlessly... So there is something wrong with Samsung implementation.
I concluded finally that it's probably worth to wait for the 2.2 release (as
jarsiv said). Hope it will not be too much delayed.
Then if we reproduce the same thing in 2.2 I'll spend more time on this
particular point.
Would have loved fring share their experience on that point (since they seems
to succeed in solving this issue) but my last tentative on their forum was a
dead end...
Another point, when I tested on the phone of my colleague, I was a little bit
disappointed with sound quality. I think that there is still improvements I can
do on that point on galaxy S (and probably all alsa based driver). I'll do more
test/debug session on galaxy when possible.
Original comment by r3gis...@gmail.com
on 17 Sep 2010 at 7:19
Yes, waiting for 2.2 will be the right way. Other way could be wasting of time.
Original comment by m...@ufritz.de
on 17 Sep 2010 at 7:32
I have take this up again :)
So is it possible that while you are calling (and put the phone on your ear)
with csipsimple it turn off the screen? Like Galaxy S does with normal GSM
calls. And it will turn on the screen when you take the phone off on your ear
(like Galaxy S does with GSM calls)
It think it uses lightness sensor or orther to find are you talking or what.
This is important feature because screen will drain your battery fast.
Original comment by jar...@gmail.com
on 27 Sep 2010 at 5:25
@galaxyS owners : could you test
http://code.google.com/p/csipsimple/downloads/detail?name=CSipSimple_0.00-13-01-
bis.apk and let me know if there is no regression regarding micro quality
during call.
I'm trying to implement something stable for other devices and also for galaxy
S in the same code... not so easy :)
However, will not solve speaker issue. This one I've to spend more time to
solve it :)
@jarsiv : latest builds (such as -13-01-bis) should already use proximity
sensor of the device if any (provided the fact the keep awake while in call
option is not manually set to true).
But as you stated, if galaxy S doesn't use a real proximity sensor it will be
harder for me to implement. But as I'm based on the same (private) API than the
official telephony application... unless they also re-implement the all
telephony application...
Original comment by r3gis...@gmail.com
on 27 Sep 2010 at 8:33
Ok, with -13-01-bis proximity sensor works now! Thanks!
I also make fast test with calls and my opinion is that now mic is choppy
(iLBC, Quality 4, mic boost 0.2, speaker boost 0.1). Can somebody Galaxy owner
test this too?
Original comment by jar...@gmail.com
on 28 Sep 2010 at 7:10
The micro quality is worse, comparing to the earlier version.
Original comment by m...@ufritz.de
on 28 Sep 2010 at 7:56
Ok, thanks for testing this version.
Sounds better than first version without anything done for alsa based driver
but indeed, not good enough compared to -12-3x series.
I have to find the good compromise since it the implementation that works
flawlessly on galaxy S introduce a regression on some other devices.
But, it's encouraging since this version give good results for other device and
not ~so bad~ for galaxy S.
So now, I have to find the good value for a parameter in my code to make it
perfect for all devices.
Stay tuned, I'll probably ask you for more tests soon :)
Original comment by r3gis...@gmail.com
on 28 Sep 2010 at 8:35
[deleted comment]
Forgot, for reference, 0.00-13-02 should be better. Tested on a real device -it
also fix speaker issue-.
In next release default values will be set for mic amplification and speaker
amplification : issue 259
Still not able to change in call volume : issue 254 .
I guess that I can remove galaxy i9000 from this issue. -> Other problem with
this device are now treated in other specific issues.
Original comment by r3gis...@gmail.com
on 3 Oct 2010 at 9:22
Yes galaxy can be removed. Version 0.00-13-02 is working well. Thank you!
Original comment by m...@ufritz.de
on 3 Oct 2010 at 4:02
It seems, that something changes on the new 2.2 update for the Galaxy I9000.
The sound issues are under development, so we can hope, the the 2.2 Froyo
firmware will behave like the other phones. Some discussion and debugging
information of this device and the sound stack:
http://forum.xda-developers.com/showthread.php?t=721069
http://forum.xda-developers.com/showthread.php?t=764548&page=2
Original comment by m...@ufritz.de
on 15 Oct 2010 at 9:01
Sounds good news for Galaxy I9000. I'm waiting for the 2.2 upgrade from samsung
on my friend device to see if it's better. (Could be a good thing to remove my
hack and behaves as it should be done regarding the api - it will be more
efficient).
As for the global bug, removed i7500 from the list of devices with problems
(tested today on another colleague device and that's fine right now :) ).
Original comment by r3gis...@gmail.com
on 18 Oct 2010 at 9:15
Original issue reported on code.google.com by
mcampbel...@gmail.com
on 14 Jul 2010 at 1:26