[Home]

Summary:ASTERISK-05295: Chanspy ruin the voice quality if any of the 3 party is using IAX
Reporter:Laszlo Tugyi (szer)Labels:
Date Opened:2005-10-14 04:26:33Date Closed:2011-06-07 14:10:29
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:I've faced a problem, which I can't see any obvious solution for. When I try to listen a conversation between A and B we can hear only inaudible noises at the end connected to the chan-spy. This is only a problem if either A or B, or the spy is softphone (IAX2 based). If i switch on the jitterbuffer and forcejitterbuffer it`s make it clear but the spying party has got a 500ms delay while the others ~80-100. Here are the tested calls

Caller    Callee    Spy       Works     With jitterbuffer settings
IAX       IAX       IAX       no        yes
IAX       SIP       IAX       no        yes
IAX       ZAP       IAX       yes       yes
IAX       SIP       SIP       yes       yes

If I try to monitor the call the state are the same like without the jitterbuffer settings.

I use IAXComm client and some grandstream hardphones.

Here is my iax.conf:

[general]
delayreject=yes
amaflags=billing
accountcode=iax
bandwidth=low
disallow=all
allow=ilbc
allow=speex
allow=gsm
allow=alaw
allow=ulaw
jitterbuffer=yes
forcejitterbuffer=yes
;dropcount=2
maxjitterbuffer=200
maxexcessbuffer=60
;minexcessbuffer=10
;jittershrinkrate=1
trunkfreq=20                    ; How frequently to send trunk msgs (in ms)
trunktimestamps=yes
tos=lowdelay
mailboxdetail=yes
regcontext=iaxreg
codecpriority=host
autokill=yes


****** ADDITIONAL INFORMATION ******

Linux-2.6.13 kernel on AMD Athlon XP 1800+ . Ubuntu-5.04
Comments:By: Kevin P. Fleming (kpfleming) 2005-10-31 18:39:49.000-0600

Some rather significant changes were made to the spy infrastructure in the past week; please update to CVS HEAD and re-test.

By: Laszlo Tugyi (szer) 2005-11-02 10:41:42.000-0600

With the new * the delay is normal, and the spy used with the monitor is working great. Thanks for the great job.

By: Laszlo Tugyi (szer) 2005-11-02 10:59:35.000-0600

Theres a problem with it again. If the force jitterbuffer is off the voice quality is bad(Not as bad with the previous version).
Caller|Callee|Spy|Works|With jitterbuffer
IAX   | IAX  |IAX| no  |yes
IAX   | SIP  |IAX| no  |yes
SIP   | ZAP  |IAX| no  |yes
ZAP   | IAX  |IAX| no  |yes

By: Kevin P. Fleming (kpfleming) 2005-11-07 22:18:05.000-0600

That doesn't make any sense; you say if 'force jitterbuffer is off' but then your table shows the jitterbuffer being on causing the problem.

Which is it?

By: Laszlo Tugyi (szer) 2005-11-08 00:03:29.000-0600

Sorry. I mean the table:
1. collumn : what kind of client used in the callee side
2. collumn : what kind of client used in the caller side
3. collumn : Is the connection works without force jitterbuffer
4. collumn : Is the connection works with force jitterbuffer
So i mean, if the force jitterbuffer is on the connection is works with all of the clients, but if its off the chanspy makes the voice discursive while i monitor the call. I try to be more clear in the next time.

By: Kevin P. Fleming (kpfleming) 2005-11-08 21:04:14.000-0600

I cannot reproduce this problem; I called from a Zap phone out an IAX channel, and then used Chanspy from a SIP phone. I also called from the SIP phone out the IAX channel and spied from the Zap channel. In both cases it worked as expected.

There are no jitterbuffer related settings on either end of these test IAX calls; only the defaults are in use.



By: Laszlo Tugyi (szer) 2005-11-09 09:54:44.000-0600

Did You monitor the calls? Try to monitor an IAX-IAX call and start to spy on it.

By: Kevin P. Fleming (kpfleming) 2005-11-10 22:33:47.000-0600

Your bug report has never previously mentioned using Monitor() at all. Please describe your _exact_ scenario, or we are wasting our time trying to reproduce your problem.

By: Kevin P. Fleming (kpfleming) 2005-12-12 19:44:45.000-0600

Suspended due to lack of feedback.