<div>Hi,</div>
<div> </div>
<div>I have a problem, which I am trying to resolve, but can not exactly figure out where it is going wrong.</div>
<div> </div>
<div>I have an inbound call, this call comes in via SIP and uses inband dtmf, at the begining of the dialplan I enable dtmf detection using spandsp_start_dtmf. this works fine, and my Lua recognizes digits correctly.</div>
<div>The tricky part is that I bridge this call in Lua using an api call "originate", this call is forwarded to the same switch, and is picked up by another Lua script. </div>
<div>This script, is waiting for a custom event, to end the lua, and is bridged with a 3rd call. so the 1st and 3rd call can hear each other. This 3rd call is initiated asynchronosly by an esl server. (this al works fine, and is not 'the issue'..)</div>
<div> </div>
<div>The problem is the receiving end (3rd leg) is receiving the DTMF pressed by the 1st leg twice. When I eavesdrop the 2nd leg, i only hear the dtmf once, when i eavesdrop the 3rd leg, I can hear the dtmf twice. So it is somewhere generated along the way.</div>
<div> </div>
<div>I have tryed several different settings, using stop_dtmf_generate on different legs, but can not seem to diable this regeneration of this extra dtmf.</div>
<div> </div>
<div>Any help would be appreciated as this is really causing issues on my side, </div>
<div> </div>
<div>ps, I know this '3rd leg' principle might look a bit weird, but cannot be avoided, </div>
<div>ps2. When my inbound call comes in using rfc2833, everything works perfectly.</div>
<div> </div>
<div> </div>
<div>Best regards,</div>
<div>Michael Lutz.</div>
<div> </div>
<div> </div>
<div> </div>
<div> </div>