«

I plan to implement an IVR solution using VoIP.

A VoIP gateway/router connects my telco voice circuits to the Ethernet
(G.711) and my app server is simply a Linux box. The app answers call,
plays an audio prompt, receives a touchtone DTMF reply, and then streams
the selected pre-recorded audio program.

QUESTION: Does VoIP have DTMF decoding built-in to the protocol?
Or would my app server have to decode the raw PCM digital audio in
order to detect the touch tones? Are there routines to do this already?

Any advice appreciated. THANKS!

John M. wrote:
> QUESTION: Does VoIP have DTMF decoding built-in to the protocol?
> Or would my app server have to decode the raw PCM digital audio in
> order to detect the touch tones? Are there routines to do this
> already?

I think that softwares such as Gnophone and Linphone has a C-library which
handles DTMF decoding. Maybe they can inspire you.

--
Mvh / Best regards,
Jack, Copenhagen

The email address is for real.

(John M.) wrote in
news: om:

> A VoIP gateway/router connects my telco voice circuits to the Ethernet
> (G.711) and my app server is simply a Linux box. The app answers call,
> plays an audio prompt, receives a touchtone DTMF reply, and then streams
> the selected pre-recorded audio program.
>
> QUESTION: Does VoIP have DTMF decoding built-in to the protocol?

Yes, I know for certain that H.323 has several ways to send DTMF signals to
endpoints (be it inband, as RFC<mumble> data or as H.245 UserInformation
signals). The problem is what is your gateway doing. If it uses inband
DTMF, you're in some sort of trouble, because now the endpoint has to
decode the DTMF from the media stream, otherwise it's easy.

--
Andreas

DTMF decoding (using SIP) can be inband, using RFC2833 or SIP-info (maybe
more).
My "Handytone ATA-286" adapter and BudgetPhone does this
(http://www.telappliant.com).

I have made a simple VoIP server program, that receives DTMF tones as
RFC2833 or SIP-info, and answers back with inband (g711) sounds.
Was easy to do, as the decoding was just to read data passing by and not by
decoding sound patterns.

--
Christen Fihl
http://HSPascal.fihl.net