VoIP Mailing List Archives

[helmut.kuper at ewetel.de]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

today we found a DTMF problem in an older FS build. We weren't able to
pass "*" and "#" to PSTN. FS sents those chars as "0". I fixed that on
my own. After that I searched in build 11055M to if it was already fixed
in trunk.

I found it was fixed in parts. I found that in INFO messages of
application type "dtmf-relay" the rfc2833 DTMF events "*#ABCD" are now
handled correct, but "EF" are still ignored. In INFO messages of
application type "dtmf" EVENTS 0-9 are considered. The others are still
ignored resp. changed to "0".


So I would recommend to replace the code beginning in line 3535

<<<<<<<<<<<<<<<<<<
if (*signal_ptr && (*signal_ptr == '*' || *signal_ptr == '#' ||
*signal_ptr == 'A' || *signal_ptr == 'B' || *signal_ptr == 'C' ||
*signal_ptr == 'D')) {
dtmf.digit = *signal_ptr;
} else {
tmp = atoi(signal_ptr);
dtmf.digit =
switch_rfc2833_to_char(tmp);
}
<<<<<<<<<<<<<<<<<<


with this two lines:


tmp=switch_char_to_rfc2833(*signal_ptr);
dtmf.digit = switch_rfc2833_to_char(tmp);


This will handle all rfc2833 dtmf events in dtmf-relay INFO messages
correctly. This is tested by me with snom and AVM clients.


Further replace the line 3556

<<<<<<<<<<<<<<
int tmp = atoi(sip->sip_payload->pl_data);
<<<<<<<<<<<<<<

with this line:
int tmp=switch_char_to_rfc2833(*(sip->sip_payload->pl_data));

That should handle all rfc2833 DTMF-Events in dtmf INFO messages.
This last patch isn't tested by me, cause I've no phone which genereates
INFO messages of type applicatin/dtmf...

best regards and a late "happy new year" to all of u!
Helmut
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (MingW32)

iEYEARECAAYFAkll54QACgkQ4tZeNddg3dxJgwCfULx8X9hw6xDd1q8r/Iih5unx
c6EAn1T/7GK56pLLrMi140x50HLnO6i1
=3ENR
-----END PGP SIGNATURE-----

_______________________________________________
Freeswitch-users mailing list
Freeswitch-users@lists.freeswitch.org
http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
http://www.freeswitch.org

[mike at jerris.com]

This workaround was added to address phones that specifically send
info dtmf incorrectly. Do you have a specific device that is not
working with 1.0.2? If so, can you please show the exact packet it is
sending.

Mike

On Jan 8, 2009, at 6:46 AM, Helmut Kuper wrote:

[helmut.kuper at ewetel.de]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Mike,

I guess the workaround refers to INFO messages with "dtmf" mime type
instead of "dtmf-relay"?

With the actual trunk (11090) it works.

DTMF event F is still converted to zero (0). Im not sure if this event
will ever transmitted via INFO.

If you use FS switch_utils function "SWITCH_DECLARE(unsigned char)
switch_char_to_rfc2833(char key)" instead of the big if-statement
(checking for *,#,A,...) befor calling "switch_rfc2833_to_char()"
function you can clean up the code a bit and handle all rfc2388 DTMF events.

The INFO message my Snom and AVM phones send, looks like this:

recv 506 bytes from udp/[85.16.245.220]:1024 at 14:59:40.229120:
------------------------------------------------------------------------
INFO sip:mod_sofia@85.16.246.6:5060;transport=udp SIP/2.0
Via: SIP/2.0/UDP 85.16.245.220:1024;branch=z9hG4bK-0a38f681korf;rport
From: "HK@FreeSWITCH" <sip:2709@85.16.246.6>;tag=v0fmpqh4nz
To: <sip:xxxxxxxx@xxxxxxx;user=phone>;tag=m0NcZ8BDZ8pcD
Call-ID: 3c26a924030f-p23y096nom3i
CSeq: 3 INFO
Max-Forwards: 70
Contact: <sip:2709@85.16.245.220:1024;line=3tzvc1jp>;reg-id=1
User-Agent: snom370/7.3.12
Content-Type: application/dtmf-relay
Content-Length: 27

Signal=*
Duration=160
------------------------------------------------------------------------

regards
Helmut

Am 08.01.2009 15:09, schrieb Michael Jerris:

[brian at freeswitch.org]

I just opened a bug with Snom on this one.. it should NEVER send a * or # in the signal line. But we do work around this but its wrong.

/b

On Jan 8, 2009, at 9:32 AM, Helmut Kuper wrote:

[helmut.kuper at ewetel.de]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Brian,

sorry, got lost ... It's a bug in Snom and AVM's phones?

Maybe we misunderstood. RFC 2833 emtioned in chapter 3.10 following DTMF
Events:

3.10 DTMF Events

Table 1 summarizes the DTMF-related named events within the
telephone-event payload format.

Event encoding (decimal)
_________________________
0--9 0--9
* 10
# 11
A--D 12--15
Flash 16

Table 1: DTMF named events

Most phones in germany are able to send * and # DTMF events and
Phone-applications like voiceboxes or IVRs need those Events.

Therefor I don't think it's a bug in Snom or AVM ...

But maybe you mean Snom and AVM should send "Signal=10" instead of
"Signal=*" ?

Unfortunately I can't find an rfc which describes DTMF over INFO
messages to find out what the right way is.

regards
helmut

Am 08.01.2009 16:43, schrieb Brian West:

[kristian.kielhofner at...]

On 1/8/09, Helmut Kuper <helmut.kuper@ewetel.de> wrote:

[brian at freeswitch.org]

Snom has already responded to my issue and are going to be providing
me a firmware for testing this tomorrow.. its still going to default
to the WRONG way.. but has a toggle to turn it to the right way.

/b

On Jan 8, 2009, at 3:23 PM, Kristian Kielhofner wrote:

[mike at jerris.com]

Update on this... the latest of anything to speak of in the standards process is that they are killing dtmf-relay. The last draft on it was:

http://tools.ietf.org/html/draft-kaplan-sipping-dtmf-package-00


Which says snom is sending it right. We will continue to accept both ways but need to think a bit about how we send. Should we match the draft or the cisco way (cisco by the way was on record saying that info dtmf-relay should die in favor of new specs such as kpml).


If anyone wants a good read of what the ietf has standardized in place of info based dtmf, see http://tools.ietf.org/html/rfc4730


Mike

On Jan 8, 2009, at 6:38 PM, Brian West wrote: