<HTML><HEAD></HEAD>
<BODY>
<DIV id=idOWAReplyText99318 dir=ltr>
<DIV dir=ltr><FONT color=#000000><FONT face=Arial size=2>Joshua Colp wrote:<BR></FONT></FONT></DIV></DIV>
<DIV dir=ltr><FONT face=Arial size=2>> Using the newest version of the speech API once a DTMF digit is received <BR>> then Asterisk will stop doing speech recognition and switch strictly to <BR>> DTMF in itself. Once this happens the SRE is no longer involved as it is <BR>> not the one who detects or handles the DTMF. If everything is properly <BR>> setup then DTMF tones should never reach the SRE. In the case where <BR>> you've got DTMF going inband and you're not setup to process it inband, <BR>> then it would spill over to the engine but that's the only case I can <BR>> think of.<BR><BR>Thanks for the explanation. So this means there is no point loading a DTMF grammar in the SRE?</FONT></DIV>
<DIV dir=ltr><FONT face=Arial size=2></FONT> </DIV>
<DIV dir=ltr><FONT face=Arial size=2>I'm actually driving this via AGI. For my current DTMF-based IVR, I use "STREAM FILE" and "WAIT FOR DIGIT", which both allow me to get the digits(s) pressed. For the ASR version, I'm using "EXEC" to call SpeechBackground, but I don't think that would return the digit as a result. As the application logic is quite complex, I don't want to handle DTMF in the dialplan in the traditional Asterisk IVR manner, so it looks as though I'll have to abandon trying to allow both speech and DTMF at the user's choice, and the application will have to be in one mode or the other.</FONT></DIV>
<DIV dir=ltr><FONT face=Arial size=2></FONT> </DIV>
<DIV dir=ltr><FONT face=Arial size=2>Any suggestions?</FONT></DIV>
<DIV dir=ltr><FONT face=Arial size=2></FONT> </DIV>
<DIV dir=ltr><FONT face=Arial size=2>David</FONT></DIV>
<DIV dir=ltr><FONT face=Arial size=2></FONT> </DIV>
<DIV dir=ltr><FONT face=Arial size=2></FONT> </DIV></BODY></HTML>