[asterisk-bugs] [Asterisk 0013053]: Called Party's inband DTMF removed almost entirely with overlapdial and non-native bridging

[Asterisk Bug Tracker]

A NOTE has been added to this issue. 
====================================================================== 
http://bugs.digium.com/view.php?id=13053 
====================================================================== 
Reported By:                Seb7
Assigned To:                jpeeler
====================================================================== 
Project:                    Asterisk
Issue ID:                   13053
Category:                   Channels/chan_dahdi
Reproducibility:            always
Severity:                   minor
Priority:                   normal
Status:                     assigned
Asterisk Version:           1.4.19 
SVN Branch (only for SVN checkouts, not tarball releases): N/A 
SVN Revision (number only!):  
Disclaimer on File?:        N/A 
Request Review:              
====================================================================== 
Date Submitted:             2008-07-10 09:50 CDT
Last Modified:              2008-12-12 11:55 CST
====================================================================== 
Summary:                    Called Party's inband DTMF removed almost entirely
with overlapdial and non-native bridging
Description: 
If you are using inband DTMF (tested on two Zap channels), and two bridged
channels are not bridged natively (e.g. because you are using MixMonitor),
and the incoming call used overlap dialing, and the Called Party tries
sending DTMF, you hear a blip and perhaps a few milliseconds of tone, but
not enough for the DTMF to be recognised by an application on the Calling
Party side. (Asterisk does not recognize the DTMF either). It is being
trimmed to within a millisecond of its life. Note, that the Calling Party's
DTMF is OK.
====================================================================== 

---------------------------------------------------------------------- 
 (0096329) Seb7 (reporter) - 2008-12-12 11:55
 http://bugs.digium.com/view.php?id=13053#c96329 
---------------------------------------------------------------------- 
russell wrote: 'The inband DTMF detector listens for DTMF and mutes it as
soon as it detects the digit.'
What you describe does seem like it might be what is happening (with some
leakage), except that the Asterisk core does not detect the DTMF: you can
see from my log that DTMF is not being detected - I was pressing DTMF keys
on my phone after I answered the call, and you can see that I have DTMF
logging switched on. (This is in contrast to jpeeler's log which shows that
DTMF is being detected, and looks like what I saw with MixMonitor but
without overlapdial on (because we can see that VLDTMF is not in use as the
tones are all 0 or 100 ms).) The fact that in my log DTMF was not being
logged is probably the reason the DTMF is not being regenerated either (or
another symptom of the same issue). So why would asterisk be muting the
DTMF but not really detecting it, or the detection not being passed all the
way though? Could it be an interaction with MixMonitor?

I do remember an issue where I complained on #asterisk-bugs that VLDTMF
didn't work with MixMonitor, [i.e. non-native bridging?], and DTMF was
converted into fixed durations (search for Seb7 in
http://ibot.rikers.org/%23asterisk-bugs/20080708.html.gz), but I think what
you have described explains this. Might be worth reading my thoughts /
findings at the time though. It looks like I didn't file this as a bug
though since no-one said it was one (although they didn't say it wasn't
either). 

Issue History 
Date Modified    Username       Field                    Change               
====================================================================== 
2008-12-12 11:55 Seb7           Note Added: 0096329                          
======================================================================