List Info

Thread: Request for Comments: RTP out-of-band dtmf signalling for connections between asteris




Request for Comments: RTP out-of-band dtmf signalling for connections between asteris
user name
2007-12-19 02:58:05
The following issue arrose out of the bug report at http://bugs.digium.com/view.php?id=11489.

I was asked by Eliel to post my thoughts to the developer list:

Beginning with asterisk-1.4.15 a bug concerning rfc2833 oob dtmf signalling on sip trunks was closed.
Unfortunately this fix opened a new issue:

Cisco carrier components (bt) use x-nse in the sdp message rather than the
telephone-event (as defined in rfc2833) - although x-nse was meant to be a complement.
Concerning the interoperability between asterisk and those components this means
that rtp out-of-band signalling does not work anymore.

I wrote a patch (for chan_sip) that, when dtmfmode=rfc2833 was set in the sip.conf, asterisk would use
rfc2833 even if the other side was not signalling this capability (resp. only signalled x-nse).

Eliel didn't feel comfortable about this as it is more of a work-around than a propper fix.

My question is:
Which way of making the two systems interoperable at this level would
be fine to digium/the developer community (and would be most likely to be
accepted into main stream code base):

1) modifying the code so that the existance of 'x-nse' in the sdp body is accepted as if
'telephone-event' was there
2) adding a config switch to sip.conf that would allow the admin to force using rfc2833 on
a sip trunk even if the other side doesn't signal rfc2833 capability

Thanks for you comments (and if I overlooked the fact that someone has already
solved this, please let me know).

-Andreas



Re: Request for Comments: RTP out-of-band dtmf signalling for connections between ast
country flaguser name
United States
2007-12-19 09:02:21
Andreas Brodmann wrote:
> 1) modifying the code so that the existance of 'x-nse'
in the sdp body
> is accepted as if
> 'telephone-event' was there

I think that this option would be fine, provided that we can
find some
documentation that verifies that "x-nse" is in
fact the same thing.  From
searching around, it looks like this has been used by Cisco
stuff for quite a
while, and it just happened to always work before the recent
change to not
include rfc2833 in the SDP if it was not offered to us in
the initial INVITE.

-- 
Russell Bryant
Senior Software Engineer
Open Source Team Lead
Digium, Inc.

_______________________________________________
--Bandwidth and Colocation Provided by http://www.api-digital.c
om--

asterisk-dev mailing list
To UNSUBSCRIBE or update options visit:
   http://lists.digium.com/mailman/listinfo/asterisk-dev

Re: Request for Comments: RTP out-of-band dtmf signalling for connections between ast
user name
2007-12-20 09:22:08
2007/12/19, Russell Bryant < russelldigium.com">russelldigium.com>:
Andreas Brodmann wrote:
>; 1) modifying the code so that the existance of 'x-nse' in the sdp body
> is accepted as if
> 'telephone-event&#39; was there

I think that this option would be fine, provided that we can find some
documentation that verifies that "x-nse" is in fact the same thing.&nbsp; From
searching around, it looks like this has been used by Cisco stuff for quite a
while, and it just happened to always work before the recent change to not
include rfc2833 in the SDP if it was not offered to us in the initial INVITE.

The X-NSE is a supplement to RFC2833 and uses the same format. So if the X-NSE occurs in
an SDP message this means that the other end supports rfc2833 plus the X-NSE.

Explanation by cisco: http://www.iana.org/assignments/media-types/audio/vnd.cisco.nse

I have attached a tiny patch to the bug report at
http://bugs.digium.com/view.php?id=11489

The patch applies against main/rtp.c and adds a new mimetype having the same
rtpPayloadType and type as 'telephone-event&#39; but has 'X-NSE' as subtype.

-Andreas

[1-3]

about | contact  Other archives ( Real Estate discussion Medical topics )