[Home]

Summary:ASTERISK-01555: SIP - IAX2 - SIP Voice Quality
Reporter:rburkholder (rburkholder)Labels:
Date Opened:2004-05-07 07:59:44Date Closed:2011-06-07 14:05:08
Priority:MajorRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:( 0) c.html
Description:I think we are getting closer and closer to a resolution, but I still am having a problem.  Here is connection:

PRI -> Cisco 3662 GW -> SIP over T1 -> RH Fedora c1 PIII/700 -> IAX2 Trunk over LAN -> RH 9 PII/233 -> SIP over LAN -> Cisco 7940

g.711 throughout

RTP timestamping isn't quite perfect on the SIP connection to the 7940.  In tcpdump/ethereal traces, I get two different type of problems (third last column is difference in sequence numbers, second last column is difference in packet sample count, last column is time difference in sec.):

Example 1 with 0 in last column:

252 4.57681 10.1.1.12 -> 10.1.6.2 Seq= 21523 , Time= 18728 1 160 0.019931
255 4.617139 10.1.1.12 -> 10.1.6.2 Seq= 21524 , Time= 18888 1 160 0.040329
256 4.617262 10.1.1.12 -> 10.1.6.2 Seq= 21525 , Time= 18888 1 0 0.000123
258 4.636805 10.1.1.12 -> 10.1.6.2 Seq= 21526 , Time= 19048 1 160 0.019543
261 4.677121 10.1.1.12 -> 10.1.6.2 Seq= 21527 , Time= 19208 1 160 0.040316
262 4.677244 10.1.1.12 -> 10.1.6.2 Seq= 21528 , Time= 19208 1 0 0.000123
264 4.696782 10.1.1.12 -> 10.1.6.2 Seq= 21529 , Time= 19368 1 160 0.019538
266 4.716937 10.1.1.12 -> 10.1.6.2 Seq= 21530 , Time= 19528 1 160 0.020155

Example 2 with a 168 in last column rather than 160:

214 4.196876 10.1.1.12 -> 10.1.6.2 Seq= 21504 , Time= 15680 1 160 0.019995
216 4.216861 10.1.1.12 -> 10.1.6.2 Seq= 21505 , Time= 15840 1 160 0.019985
218 4.236855 10.1.1.12 -> 10.1.6.2 Seq= 21506 , Time= 16008 1 168 0.019994
220 4.256935 10.1.1.12 -> 10.1.6.2 Seq= 21507 , Time= 16168 1 160 0.02008
222 4.276853 10.1.1.12 -> 10.1.6.2 Seq= 21508 , Time= 16328 1 160 0.019918

When these happen, sound gets silenced and it may take several seconds for the sound to come back again.
Comments:By: Mark Spencer (markster) 2004-05-07 09:25:59

The bogosity meter reads extremely high, once again.

This issues has been addressed already, including for iax2 trunking.  I'll be happy to take a look if you find me on IRC, *but* before that, you should be ABSOLUTELY POSITIVELY SURE that you have cvs HEAD at any Asterisk location.
By: rburkholder (rburkholder) 2004-05-07 09:29:05

fedora machine:
srvr16*CLI> show version
Asterisk CVS-HEAD-05/06/04-19:15:59 built by root@srvr16.oneunified.net on a i686 running Linux
srvr16*CLI>

rh9 machine:
netmon*CLI> show version
Asterisk CVS-HEAD-05/06/04-18:35:03 built by root@netmon.oneunified.net on a i686 running Linux
By: Mark Spencer (markster) 2004-05-07 09:35:56

Alright just find me on IRC (kram)
By: rburkholder (rburkholder) 2004-05-07 09:40:51

Ok.  A change to rtp.c wasn't over-written with new CVS.  File is now updated properly, and issue appears to have disappeared.