[Standards] Another Jingle implementation
paulrw at codian.com
Fri Feb 8 11:17:37 UTC 2008
Peter Saint-Andre wrote:
> Paul Witty wrote:
>> Peter Saint-Andre wrote:
>>>> I already have support for
>>>> in-band DTMF, and, once I've written the stuff to negotiate it, will
>>>> support RFC 4733 DTMF in the content channel. I'll also implement DTMF
>>>> in the signalling channel.
>>> I see two options:
>>> 1. DTMF over XMPP (XEP-0181) ("in-band/signalling")
>>> 2. DTMF over RTP (RFC 4733) ("out-of-band/media")
>>> Do you see a third option?
>> DTMF actually as tones in the audio. Old-fashioned, but valid enough
>> for G.711, and probably some other codecs as well. Sometimes needed if
>> going to a SIP to POTS gateway, and not a bad fallback if the far end
>> doesn't offer support for XEP-0181 or RFC 4733.
> Ah, OK, I got confused by terminology. Typically people call the
> "signalling channel" something like SIP or XMPP (where you negotiate the
> media session) and the "media channel" something like RTP (the transport
> for the media). You're talking about sending DTMF as tones in the media
> itself. Yes, that's old-fashioned, but it probably gets the job done. :)
I'd consider DTMF tones in the audio to be the only in-band method, with
RFC 4733 being out-of-band as it is the same channel but different band
as the media, and XEP-0181 being out-of-band as it is in a different
channel to the media. I may have a slightly video-conferencing view of
the world though.
For a bit of a laugh, I've also implemented DTMF through chat messages
(as well as a very basic chat-based UI to connect to pre-defined
destinations through the gateway), in order to let released versions of
Google Talk send messages.
More information about the Standards