[Freeswitch-users] tone_detect vs. ignore_early_mediaANDmod_fifo

Jeroen C. van Gelderen slim at thegreek.com
Wed Oct 13 03:23:22 PDT 2010 

I was just going to respond to your previous mail: I have commented out the
pause/resume statements and this solved my problem.

I will try the latest git when I get back in office. Thank you.

Cheers,
-Slim
--
Jeroen C. "Slim" van Gelderen

-----Original Message-----
From: freeswitch-users-bounces at lists.freeswitch.org
[mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of Anthony
Minessale
Sent: Tuesday, October 12, 2010 12:34
To: FreeSWITCH Users Help
Subject: Re: [Freeswitch-users] tone_detect vs.
ignore_early_mediaANDmod_fifo

try latest GIT

On Mon, Oct 11, 2010 at 4:00 PM, Anthony Minessale
<anthony.minessale at gmail.com> wrote:
> This was done to stop recordings and other things on callers waiting in a
queue.
> We may have to change it to be configurable.
>
>
> On Mon, Oct 11, 2010 at 6:51 AM, Jeroen C. van Gelderen
> <slim at thegreek.com> wrote:
>>
>>
>> I have an inkling that the following commit made between 1.0.2 and 1.0.3
>> might have something to do with this:
>>
>>
>>
>>    * mod_fifo: pause media bugs while not in a bridge (r:11466,11490)
>>
>>
>>
>>
http://lists.freeswitch.org/pipermail/freeswitch-svn/2009-January/009566..ht
ml
>>
>>
http://lists.freeswitch.org/pipermail/freeswitch-svn/2009-January/009589..ht
ml
>>
>>
>>
>> Can anyone comment on the how and why? Is there any way to reconcile this
>> with use of tone_detect?
>>
>> Cheers,
>> -Slim
>> --
>> Jeroen C. "Slim" van Gelderen
>> Olympic Sports Data Services
>> Email: slim at thegreek.com
>> Phone: +1 876 953 6182 x128
>>
>> ________________________________
>>
>> From: freeswitch-users-bounces at lists.freeswitch.org
>> [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of
Jeroen
>> C. van Gelderen
>> Sent: Monday, October 11, 2010 05:31
>> To: 'FreeSWITCH Users Help'
>> Subject: Re: [Freeswitch-users] tone_detect vs. ignore_early_media
>> ANDmod_fifo
>>
>>
>>
>>
>>
>> Hmm
 the plot thickens

>>
>>
>>
>> When I bridge my FXO port to the SIP-GSM gateway directly (i.e. without
>> using mod_fifo) I don’t seem to need monitor_early_media_fail. Using
>> tone_detect on the A leg works fine when “ignore_early_media=true” is
used
>> on the B leg:
>>
>>
>>
>> The following dialplan excerpt WORKS (i.e. FXO hang-ups are detected at
all
>> stages by tone_detect):
>>
>>
>>
>>     <extension name="pan_125">
>>
>>       <condition field="${freetdm_span_name}"
>> expression="^span_fxo_helpdesk$">
>>
>>         <!-- Detect busy tone from Panasonic -->
>>
>>         <action application="tone_detect" data="busy-pan 350,440 r 0
hangup
>> undef 2"/>
>>
>>
>>
>>         <action application="answer"/>
>>
>>
>>
>>         <action application="set" data="continue_on_fail=true"/>
>>
>>         <action application="set" data="instant_ringback=true"/>
>>
>>         <action application="set"
data="ringback=%(1000,1500,440.0,480.0)"/>
>>
>>
>>
>>         <action application="bridge"
>>
data="{ignore_early_media=true}sofia/internal/1??????76??@192.168.3.11:5060"
/>
>>
>>         <action application="bridge"
>>
data="{ignore_early_media=true}sofia/internal/1??????77??@192.168.3.11:5060"
/>
>>
>>       </condition>
>>
>>     </extension>
>>
>>
>>
>> The problem seems to occur only when mod_fifo is added to the mix:
>>
>>
>>
>>     <extension name="test2">
>>
>>       <condition field="${freetdm_span_name}"
expression="^span_fxo_test$">
>>
>>         <!-- Detect busy tone from Panasonic -->
>>
>>         <action application="tone_detect" data="busy-pan 350,440 r 0
hangup
>> undef 2"/>
>>
>>
>>
>>         <action application="answer"/>
>>
>>
>>
>>         <action application="playback" data="thankyou.wav"/>
>>
>>         <action application="set" data="fifo_music=$${hold_music}"/>
>>
>>         <action application="fifo" data="fifo_helpdesk in"/>
>>
>>       </condition>
>>
>>     </extension>
>>
>>
>>
>> In all cases:
>>
>> -          tone_detect works BEFORE the call is handed to mod_fifo (i.e.
>> during playback)
>>
>> -          tone_detect works AFTER the call is established by mod_fifo
and
>> audio is being exchanged between A and B leg.
>>
>>
>>
>> IF the A leg (with tone_detect enabled) is handed to mod_fifo AND there
are
>> no agents in the fifo THEN tone_detect does not work. Mod_fifo simply
plays
>> MOH to the A leg perpetually.
>>
>>
>>
>> IF the A leg (with tone_detect enabled) is handed to mod_fifo AND
mod_fifo
>> places an outbound call with “ignore_early_media=true” THEN tone_detect
does
>> not work on the A leg during the early media phase on leg B.
>>
>>
>>
>> IF the A leg (with tone_detect enabled) is handed to mod_fifo AND
mod_fifo
>> places an outbound call with “ignore_early_media=false” THEN tone_detect
>> does work. I guess this is because audio is being exchanged between A and
B
>> legs.
>>
>>
>>
>> Any idea what would cause the tone_detect to be “suspended” when mod_fifo
is
>> in the mix?
>>
>> Cheers,
>> -Slim
>> --
>> Jeroen C. "Slim" van Gelderen
>>
>> ________________________________
>>
>> From: freeswitch-users-bounces at lists.freeswitch.org
>> [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of
Jeroen
>> C. van Gelderen
>> Sent: Monday, October 11, 2010 03:13
>> To: 'FreeSWITCH Users Help'
>> Subject: Re: [Freeswitch-users] tone_detect vs. ignore_early_media (per
>> leg?)
>>
>>
>>
>> Thank you for the quick response.
>>
>>
>>
>> It looks like monitor_early_media_fail should do what I need (thanks for
the
>> suggestion!) but I can’t seem to make it work. See below for my
uneducated
>> best guess why.
>>
>>
>>
>> In my case the failure conditions from the GSM side are handled
out-of-band
>> by SIP. That leaves only one failure condition I need to listen for in
early
>> media (the Panasonic far-end hang up on FXO) which can successfully be
>> detected with:
>>
>>
>>
>>   <action application="tone_detect" data="busy-pan 350,440 r 0 hangup>
>>
>>
>>
>> This results in the following log entries:
>>
>>
>>
>> [DEBUG] switch_ivr_async.c:2072 Adding tone spec 350,440 index 0 hits 1
>>
>> [DEBUG] switch_core_media_bug.c:360 Attaching BUG to FreeTDM/4:1/
>>
>>
>>
>> Translating tone_detect to monitor_early_media_fail I tried the
following:
>>
>>
>>
>>   <action application="set" data="result=${fifo_member(add fifo_helpdesk
>>
{fifo_member_wait=nowait,ignore_early_media=true,monitor_early_media_fail=bu
sy-pan:1:350+440}sofia/internal/${caller_id_number}@192.168.3.11:5060)}"/>
>>
>>
>>
>> This results in the following relevant log entries:
>>
>>
>>
>> [DEBUG] switch_channel.c:2444
(sofia/internal/1876??????@192.168.3.11:5060)
>> Callstate Change RINGING -> EARLY
>>
>> [DEBUG] switch_ivr_async.c:2072 Adding tone spec 350,440 index 0 hits 1
>>
>> [DEBUG] switch_core_media_bug.c:360 Attaching BUG to
>> sofia/internal/1??????7693 at 192.168.3.11:5060
>>
>>
>>
>> So Freeswitch is listening for the right tones (tone_spec present and
>> identical to the one used in tone_detect approach) but it isn’t detecting
>> them. The obvious difference is in the BUG attachment.
>>
>>
>>
>> Is it possible that BUG isn’t listening to the right (A) leg in the case
of
>> monitor_early_media_fail? Or is this too easy? J
>>
>> Cheers,
>> -Slim
>> --
>> Jeroen C. "Slim" van Gelderen
>>
>> ________________________________
>>
>> From: freeswitch-users-bounces at lists.freeswitch.org
>> [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of
Michael
>> Collins
>> Sent: Sunday, October 10, 2010 19:01
>> To: FreeSWITCH Users Help
>> Subject: Re: [Freeswitch-users] tone_detect vs. ignore_early_media (per
>> leg?)
>>
>>
>>
>> Check out monitor_early_media_fail:
>>
http://wiki.freeswitch.org/wiki/Channel_Variables#monitor_early_media_fail
>>
>> It is a hybrid of ignoring early media and listening to early media for
>> various conditions. In your case you'll need to play around with it. In
your
>> case you need to figure out which early media scenarios count as a "fail"
>> and will cause processing to move on as if the call really did fail.
>>
>> Roll up your sleeves, you have some work to do. :)
>>
>> -MC
>>
>> On Sat, Oct 9, 2010 at 9:24 PM, Jeroen C. van Gelderen
<slim at thegreek.com>
>> wrote:
>>
>> Hi Guys,
>>
>> I have a legacy Panasonic PBX which does not support CPC/Disconnect
>> Supervision. Calls from this PBX are sent to Freeswitch by way of
>> DAHDI/FreeTDM FXO ports (on a Xorcom Astribank). I use tone_detect to
deal
>> with far-end hang-ups. This works fine:
>>
>> <!-- Detect busy tone from Panasonic -->
>> <action application="tone_detect" data="busy 350,440 r 0 hangup 4"/>
>>
>> Some calls from the Panasonic PBX are put in a FIFO and from there they
are
>> sent to a SIP-to-GSM bridge (GoIP) to be terminated on one of 3 GSM
phones
>> in a round-robin fashion.
>>
>> The GSM bridge requires me to use "ignore_early_media=true" otherwise the
>> caller will receive messages like "the number you are calling does not
>> answer". When I set "ignore_early_media=true" the FIFO correctly keeps
>> hunting for a GSM phone that is actually answered and will ignore phones
>> that are busy, no-answer or turned off. This too works fine.
>>
>> The problem occurs when the two are combined as follows:
>>
>> Panasonic PBX --FXO--> Freeswitch --SIP--> GSM-bridge
>>
>> If I enable ignore_early_media then tone_detect doesn't work UNTIL one of
>> the GSMs is answered. This is a problem when none of the GSMs are
answered
>> and the Panasonic PBX hangs up. The FXO port will stay "up" and the FIFO
>> will keep hunting until it times out.
>>
>> If I don't use "ignore_early_media=true" then tone_detect works fine but
>> then every telco message gets mistaken for an answered call and the
hunting
>> stops early.
>>
>> I tried changing this example line from my fifo.conf:
>>
>> <member timeout="30" simo="1" lag="5">
>>
{member_wait=nowait,ignore_early_media=true}sofia/internal/1??????7701 at 192.1
>> 68.3.11:5060
>> </member>
>>
>> to
>>
>> <member timeout="30" simo="1" lag="5">
>>
{member_wait=nowait}[ignore_early_media=true]sofia/internal/1??????7701 at 192.
>> 168.3.11:5060
>> </member>
>>
>> in a vain attempt to ignore early media on the SIP leg only. This doesn't
>> seem to do anything however.
>>
>> Can anyone clue me in on what I'm missing? I've snipped the relevant
>> configuration bits below. I have the feeling I'm missing something
obvious.
>>
>> Cheers,
>> -Slim
>>
>> ----8<----8<----8<----8<----8<----
>>
>> <configuration name="fifo.conf" description="FIFO Configuration">
>>  <settings>
>>    <param name="delete-all-outbound-member-on-startup" value="false"/>
>>  </settings>
>>  <fifos>
>>    <fifo name="cool_fifo@$${domain}" importance="0">
>>
>>      <member timeout="30" simo="1" lag="5">
>>
{member_wait=nowait,ignore_early_media=true}sofia/internal/1??????7701 at 192.1
>> 68.3.11:5060</member>
>>
>>      <member timeout="30" simo="1" lag="5">
>>
{member_wait=nowait,ignore_early_media=true}sofia/internal/1??????7693 at 192.1
>> 68.3.11:5060</member>
>>
>>      <member timeout="30" simo="1" lag="5">
>>
{member_wait=nowait,ignore_early_media=true}sofia/internal/1??????7698 at 192.1
>> 68.3.11:5060</member>
>>
>>    </fifo>
>>  </fifos>
>> </configuration>
>>
>> ----8<----8<----8<----8<----8<----
>>
>> <include>
>>  <context name="pbx_test">
>>
>>    <extension name="test1">
>>      <condition field="${freetdm_span_name}"
expression="^span_fxo_test$">
>>        <action application="answer"/>
>>
>>        <!-- Detect busy tone from Panasonic -->
>>        <action application="tone_detect" data="busy 350,440 r 0 hangup
4"/>
>>
>>        <action application="playback"
>> data="ivr/ivr-hold_connect_call.wav"/>
>>
>>        <action application="fifo" data="cool_fifo@$${domain} in"/>
>>
>>      </condition>
>>    </extension>
>>  </context>
>> </include>
>>
>> ----8<----8<----8<----8<----8<----
>>
>> Cheers,
>> -Slim
>> --
>> Jeroen C. "Slim" van Gelderen
>>
>>
>>
>> _______________________________________________
>> FreeSWITCH-users mailing list
>> FreeSWITCH-users at lists.freeswitch.org
>> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
>> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
>> http://www.freeswitch.org
>>
>>
>>
>> _______________________________________________
>> FreeSWITCH-users mailing list
>> FreeSWITCH-users at lists.freeswitch.org
>> http://lists.freeswitch.org/mailman/listinfo/freeswitch-users
>> UNSUBSCRIBE:http://lists.freeswitch.org/mailman/options/freeswitch-users
>> http://www.freeswitch.org
>>
>>
>
>
>
> --
> Anthony Minessale II
>
> FreeSWITCH http://www.freeswitch.org/
> ClueCon http://www.cluecon.com/
> Twitter: http://twitter.com/FreeSWITCH_wire
>
> AIM: anthm
> MSN:anthony_minessale at hotmail.com
> GTALK/JABBER/PAYPAL:anthony.minessale at gmail.com
> IRC: irc.freenode.net #freeswitch
>
> FreeSWITCH Developer Conference
> sip:888 at conference.freeswitch.org
> googletalk:conf+888 at conference.freeswitch.org
> pstn:+19193869900
>



-- 
Anthony Minessale II

FreeSWITCH http://www.freeswitch.org/
ClueCon http://www.cluecon.com/
Twitter: http://twitter.com/FreeSWITCH_wire

AIM: anthm
MSN:anthony_minessale at hotmail.com
GTALK/JABBER/PAYPAL:anthony.minessale at gmail.com
IRC: irc.freenode.net #freeswitch

FreeSWITCH Developer Conference
sip:888 at conference.freeswitch.org
googletalk:conf+888 at conference.freeswitch.org
pstn:+19193869900

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

More information about the FreeSWITCH-users mailing list