DMTF tones being mis-interpreted

[mmertens]

CUCM 6.1(2) with MGCP GWs:

I am having an intermittent issue when users are hopping onto (external) conference calls where they have to enter a PIN/conference #, and the conference bridge tells them they are hitting extra digits, ex: person enters 7165, and the bridge reports that they hit 777165 or 7111165 or 716665, etc. Call set-up/dialing is NOT a problem. (And I don't know if this is happening with other applications that require DTMF input).

My original thought was something with the H225 DTMF timer, but now I'm wondering whether it isn't a voice quality issue between the phone and voice gateway. I could picture if a single pressing of one DTMF number would produce multiple packets, and one of those packets are dropped, it may appears as though the number was pressed multiple times? Yet, it may not be a big enough voice quality problem where people complain.

Any thoughts? Any experience with this? Any input?

Thanks,

[lagos · Accepted solution]

You may be running into BugID CSCsv04751. This can impact 3rd generation, 7961,7965,7970 phones and cause DTMF to be send in the in-band audio stream. This only impacts the phone when on handset and not on speakerphone. The behavior is present in firmware load 8.4(1)SR1S and fixed in 8.4(2).

[mmertens]

Coincidently, I'm running 8-4-1SR1S for phone loads.....I'd say your solution is a strong candidate.

THANKS!

[Nicholas Matthews]

This is actually my bug :)

You can test this very easily - go on your handset and press a DTMF digit down longer than 200ms or so. This will cause in-band RTP DTMF to be generated in addition to the Skinny Keypad message.

Now, if you use the speakerphone or a headset this doesn't happen. This has to do with the monotonous beep played for both headsets and speaker phones instead of an actual DTMF that is played into the sidetone of the handset.

This is in phone loads 8.4.1 and 8.4.1 Sr 1, fixed in 8.4.2 and later, 8.4.3, etc.

hth,

nick