|Summary:||ASTERISK-01202: [request] go to fax extension when fax signal received during voice call without user interaction|
|Reporter:||Diego Ercolani (dercol)||Labels:|
|Date Opened:||2004-03-12 09:08:10.000-0600||Date Closed:||2011-06-07 14:05:04|
|Description:||When answered a fax call, asterisk complain about it telling that it can't represent 'f'.|
It would be wonderful if it would call the extension f, so is it possible to accomplish the fax reception
****** ADDITIONAL INFORMATION ******
Received fax call via AVM B1 ISDN card from a standard fax machine
|Comments:||By: zoa (zoa) 2004-03-12 09:34:47.000-0600|
the f extension already exists
By: Diego Ercolani (dercol) 2004-03-12 10:02:17.000-0600
I know there is 'f' extension but it seems isn't achieved from the DTMF routines.
Sorry for my poor explaination but I hope this session would explai you what I'm trying to obtain:
amarone*CLI> set verbose 999999999999999999999999999999999999
-- creating pipe for PLCI=0x101 msn = *
> sent ALERT_REQ PLCI = 0x101
-- Executing Wait("CAPI[contr1/515387092]/66", "1") in new stack
-- started pbx on channel (callgroup=0)!
-- Executing GotoIf("CAPI[contr1/515387092]/66", "1?diego|1:s|3") in new stack
-- Goto (CAPI,diego,1)
-- Executing Dial("CAPI[contr1/515387092]/66", "SIP/gs2&SIP/kphone2|60|t") in new stack
-- Called gs2
-- Called kphone2
-- SIP/gs2-0404 is ringing
-- SIP/gs2-0404 answered CAPI[contr1/515387092]/66
-- CAPI Answering for MSN 515387092
Mar 12 15:44:36 WARNING: rtp.c:856 ast_rtp_senddigit: Don't know how to represent 'f'
Mar 12 15:44:38 WARNING: chan_sip.c:495 retrans_pkt: Maximum retries exceeded on call email@example.com for seqno 102 (Request)
Mar 12 15:44:40 WARNING: rtp.c:856 ast_rtp_senddigit: Don't know how to represent 'f'
Mar 12 15:44:41 WARNING: chan_sip.c:495 retrans_pkt: Maximum retries exceeded on call firstname.lastname@example.org for seqno 102 (Request)
Mar 12 15:44:43 WARNING: rtp.c:856 ast_rtp_senddigit: Don't know how to represent 'f'
Mar 12 15:44:47 WARNING: rtp.c:856 ast_rtp_senddigit: Don't know how to represent 'f'
Mar 12 15:44:50 WARNING: rtp.c:856 ast_rtp_senddigit: Don't know how to represent 'f'
== Spawn extension (CAPI, diego, 1) exited non-zero on 'CAPI[contr1/515387092]/66'
-- CAPI Hangingup
> sent DISCONNECT_B3_REQ NCCI=0x10101
> sent DISCONNECT_REQ PLCI=0x101
-- removed pipe for PLCI = 0x101
As you see, asterisk complains about representing 'f' dialtone...... so it can't call extension 'f' to handle fax call while it correctly receive the fax call information!
By: zoa (zoa) 2004-03-12 10:11:27.000-0600
can you also post your dialplan (extensions.conf) ?
By: Diego Ercolani (dercol) 2004-03-12 11:00:44.000-0600
Sorry, I've understand now your answer...... infact I wasn't answer from a classic directory extension like the [demo] context in included distribution example.
The situation I would allow is:
1. person1 calls from a fax machine and begin a voice call to his partner 'person2'
2. person1 click 'start' button on the fax machine while not hanging up
3. asterisk understand via silence and presence of the 'f' tone that the call have switched to a fax call
4. asterisk switchs automatically the call to a fax handler (example 'f' extension) without the human interaction (without having to dial '#' followed by the internal of the fax machine)
edited on: 03-12-04 09:51
By: zoa (zoa) 2004-03-12 11:06:59.000-0600
Aaaaah, now i understand your question.
By: Tilghman Lesher (tilghman) 2004-03-12 11:32:31.000-0600
You have a major problem, though, in that Asterisk makes no differentiation between a fax call and a voice call. So if the presence of the fax tone forces a call into the 'fax' extension, then you could possibly have an infinite loop in the call logic. (i.e. after a call is moved to the 'fax' extension, the fax tone is sensed and the call is restarted into the 'fax' extension, ad infinitum).
The 'f' tone is only currently sensed and moved to the 'fax' extension if an appropriate application is being used at the time (one that listens for DTMF, such as Background). If you add a two second greeting to all calls incoming and play that greeting with Background, that is generally sufficient time in which to detect the fax tone and redirect to the 'fax' extension.
By: zoa (zoa) 2004-03-12 12:32:14.000-0600
but that still won't allow people to send a fax after a normal call without redialing.
The looping problem could be fixed by not looking for fax when you are already in the fax extension. (or ignoring it).
By: Tilghman Lesher (tilghman) 2004-03-12 12:58:10.000-0600
I don't think that's something you really want to play with. The tambre of some people's voices could activate the fax detection routines during a call. That's more likely to piss someone off who will insist that that's a bug (and I really can't disagree).
If you want a fax call, send the call separately.
By: zoa (zoa) 2004-03-12 13:02:48.000-0600
corydon, then why bother looking for fax sounds during a call ?
(instead of only the first X seconds)
my normal fax at home can do it like that, i can send a fax in the middle of the phone call, do you think the fax signal is so delicate that a man's voice could activate it ?
(maybe we could check for multiple detections in a small period of time ?)
By: Tilghman Lesher (tilghman) 2004-03-12 13:26:06.000-0600
We don't look for faxes while any call is bridged. We only detect it when detecting DTMF during a call (like I said before, when running the Background application or another application that accepts any DTMF).
And you _can_ send a fax in the middle of a call, as long as you do not expect Asterisk to change the endpoints of a bridge.
And no, a man's voice is generally not high enough to emulate the fax tone. However, it has happened before to certain women (not all women, only women with a certain tone and tambre).
By: heison (heison) 2004-04-06 00:29:42
what would be nice too is have T.38 implemented as part of the fax detection.
ie. when someone press send on the fax machine, it automatically switches from the existing codec (voice codec, g729 for example) to T.38
By: Brian West (bkw918) 2004-04-18 00:44:10
I agree this is kinda pointless (what about false positives?)