Summary:ASTERISK-02254: 1st second of RTP (especially indications) to sip channel is poor quality
Reporter:flydoc (flydoc)Labels:
Date Opened:2004-08-22 14:35:18Date Closed:2008-01-15 15:06:01.000-0600
Versions:Frequency of
Environment:Attachments:( 0) app_disa_c.diff
( 1) app_disa.c
Description:Ringing in extensions.conf will generate SIP 180 Ringing
which is fine.
DISA will generate tones in software. If debug is on get ast_rtp_raw_write difference is 11nn, ms is 16n several times.
Issue clears within 1st second
DISA forces ALAW. This happens with differing sip devices. and is not comfort noise relate. However we are receiving RTP at this stage


tested on almost cvs head version
Comments:By: Mark Spencer (markster) 2004-08-22 14:38:59

Obviously this isn't even remotely a MAJOR bug.

By: Mark Spencer (markster) 2004-08-22 22:34:52

Under what circumstances do you have the bad second?  In phone to phone, phone to generator (e.g. music on hold), phone to DISA or what, also what sort of devices have you used, and if the answer is "I have a 7960" then try something else and see if it still occurs.

By: Brian West (bkw918) 2004-08-22 22:54:11

"tested on almost cvs head version" doesn't cut it.. update and try again.

By: flydoc (flydoc) 2004-08-23 12:28:49

apologies about bug grading!
now running cvs head (today)
ok - as in description, caller does through to DISA in extensions.conf and hears generated dial tone, it is this that sounds very poor at the start.
This is using a mediatix 2102, on a Cisco 3660 it is fine.
I would put it down to the box, but why does * generate the ast_rtp_raw_write "difference" as per bug description?
Study of ethereal traces does not show much, other than RTP MARK is sent at start of RTP stream from Cisco, and 7th RTP packet from mediatrix.
Happy to try anything legal.

By: Brian West (bkw918) 2004-08-23 13:36:05

I know we don't support that mark stuff in rtp.. hrm

By: flydoc (flydoc) 2004-08-27 12:57:49

further works show is VAD related - works fine with VAD off on IAD
I suggest you close the Bug


By: Mark Spencer (markster) 2004-08-27 14:08:23

Try it now with latest CVS but *not* with DISA.  I've modified generators to be zap timed when a zaptel card is installed.

By: flydoc (flydoc) 2004-08-28 04:54:13

works a treat - ta

I have done DISA for you, not really a developer anymore - but appears to work

also good to hear correct dialtone if you are in UK

composing "inbound SIP channel status" bug (Up set too soon) which also impacts start of RTP streaming

By: Mark Spencer (markster) 2004-08-28 10:49:37

I'll need a "cvs diff -u" and a disclaimer to use your DISA changes, but that looks good.

By: flydoc (flydoc) 2004-08-28 11:45:30

Dislaimer fax sent

By: Mark Spencer (markster) 2004-08-28 12:34:28

Okies great, added to CVS (modified to do "dialtone" instead of ringing)...

By: flydoc (flydoc) 2004-08-28 14:58:38

OOPS sorry re ringing

However you need to call ast_playtones_start not ast_tonepair_start or we are back into USA only mode

sorry to be pedantic

By: Mark Spencer (markster) 2004-08-28 15:37:41

Okay fixed in CVS

By: Digium Subversion (svnbot) 2008-01-15 15:05:59.000-0600

Repository: asterisk
Revision: 3670

U   trunk/apps/app_disa.c

r3670 | markster | 2008-01-15 15:05:59 -0600 (Tue, 15 Jan 2008) | 2 lines

Major DISA improvements (bug ASTERISK-2254)



By: Digium Subversion (svnbot) 2008-01-15 15:06:01.000-0600

Repository: asterisk
Revision: 3672

U   trunk/apps/app_disa.c

r3672 | markster | 2008-01-15 15:06:00 -0600 (Tue, 15 Jan 2008) | 2 lines

Update DISA to be internationalized (bug ASTERISK-2254)