[Home]

Summary:ASTERISK-11902: Jitterbuffer used with Queues causes large audio delay
Reporter:John Lange (johnlange)Labels:
Date Opened:2008-04-23 11:35:18Date Closed:2011-06-07 14:03:00
Priority:MajorRegression?No
Status:Closed/CompleteComponents:Core/Jitterbuffer
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:If you configure ZAP for a fixed jitter buffer of
about 60ms the calls sent to agent phones experience a huge amount of
audio latency. I'm just estimating but I'd say its at least 600ms.

This effects only the calls coming through queues. Calls that do not go through
a Queue are fine.

Disabling the jitter buffer resolves the issue.
Comments:By: Jason Parker (jparker) 2008-04-23 13:50:20

When you say "agent", do you mean a real agent with chan_agent, or just a "standard" queue member?
By: John Lange (johnlange) 2008-04-23 16:10:39

To be honest I don't know the difference. They are not configured in agents.conf but are added dynamically via an AGI that calls "AddQueueMember".

$AGI->exec('AddQueueMember', "$queue||$pri");

Does that clarify it?
By: Jason Parker (jparker) 2008-04-23 16:16:28

Looks like regular queue members (the distinction is very important).
By: Mark Michelson (mmichelson) 2008-07-17 17:07:49

I attempted to reproduce this today using DAHDI. I added the following lines to my channel configuration for an FXS port on a Digium TDM400P.

jbenable=yes
jbforce=yes
jbimpl=fixed
jbmaxsize=60

I could not reproduce with these settings. If you had additional (or different) settings or were not using an FXS channel when this occurred, then please correct me in my settings.

If otherwise, then it may be that this issue has since been fixed.
By: John Lange (johnlange) 2008-07-21 09:27:46

In our incident, the incoming lines are a T1 PRI. This would most often be the case with queues since they are most often used with higher call volumes. I will do an upgrade to the latest version of 1.4 and test again though this will probably not happen for a week or so.
By: Leif Madsen (lmadsen) 2008-12-05 09:08:50.000-0600

johnlange: is this still an issue? Or can we close this issue?
By: John Lange (johnlange) 2008-12-10 20:35:14.000-0600

I'll try and test it again as soon as I can (hopefully next week) but as far as I know it still exists.
By: Clod Patry (junky) 2009-02-12 12:23:28.000-0600

johnlange: any feedback here?
By: John Lange (johnlange) 2009-02-12 13:54:47.000-0600

Sorry. I haven't been able to get back to that particular client and test this to see if it still exists. Probably best to just close it and I'll ask for it to be reopened should I ever get a chance to test it.
By: Joshua C. Colp (jcolp) 2009-02-12 13:55:39.000-0600

Suspended per reporter.