VoIP Mailing List Archives

[anthony.minessale at g...]

This is a common issue with analog phones even traditional answering machines suffer from it.
I'm sure you must have had an answering machine at some point that has dial tone as the message it receives.

Unless FreeSWITCH has some hint that the call has hungup it will not stop trying to complete the call.

If the other side is sending a busy tone to indicate hangup it's possible to use the tone_detect app to pick
up on the tones and abort the call.

Another thing you could do if you have unlimited inbound is explicitly answer the call in the dialplan before
you call your sip phones this will give you a more profound hangup detection but it will make every call count
even when nobody answers.



On Tue, Jan 20, 2009 at 10:46 AM, Tomás <tomasborrella@gmail.com (tomasborrella@gmail.com)> wrote:

[scott.ellis at novatex...]

A lot of the hardware has problems detecting hangups, and the OpenZap drivers are not totally sorted yet. So I would be looking at hardware/openzap issue, not FS.

Scott

Tomás wrote:

[oseslija at gmail.com]

I tried similar setup with my analog card (X100P) and I'm having same issue. Call is not hungup on the oz side once the caller ends. My telco doesn't use polarity reversal for singalling hang up so I'm guess I'm stuck to detecting busy tone from the telco side. I'll try to modify tones.conf accordingly.

Regards,
Ognjen
(sekil)

On Tue, Jan 20, 2009 at 6:05 PM, Anthony Minessale <anthony.minessale@gmail.com (anthony.minessale@gmail.com)> wrote:

[oseslija at gmail.com]

Ok, as discussed with Tony on IRC channel I followed his directions which lead to a successfull outcome (like it always does I might add .

One has to use tone_detect app in FreeSWITCH dialplan in order to check for busy tones coming from the PSTN side and if matched fire a hangup application. This is the snippet of my test dp that does the trick (from extension Local_extensions in default.xml):

<anti-action application="tone_detect" data="busy 425,0 r +100 hangup 16 4"/>
<anti-action application="bridge" data="user/${dialed_extension}@${domain_name}"/ ([email]user/$%7Bdialed_extension%7D@$%7Bdomain_name%7D%22/[/email])>

This means that FS will listen to freq of 425 Hz and wait for 4 positive detection to fire up hangup app with code 16 which is NORMAL_CLEARING (425 Hz is the freq telco here uses; for other countries I suggest getting the ITU world tones pdf file and check there):

2009-01-20 22:32:46 [DEBUG] switch_ivr_async.c:1262 tone_detect_callback() TONE busy HIT 1/4
2009-01-20 22:32:47 [DEBUG] switch_ivr_async.c:1262 tone_detect_callback() TONE busy HIT 2/4
2009-01-20 22:32:48 [DEBUG] switch_ivr_async.c:1262 tone_detect_callback() TONE busy HIT 3/4
2009-01-20 22:32:50 [DEBUG] switch_ivr_async.c:1262 tone_detect_callback() TONE busy HIT 4/4
2009-01-20 22:32:50 [DEBUG] switch_ivr_async.c:1268 tone_detect_callback() TONE busy DETECTED

2009-01-20 22:32:50 [NOTICE] mod_dptools.c:584 hangup_function() Hangup OpenZAP/1:1/1 [CS_EXECUTE] [NORMAL_CLEARING]

Regards,
Ognjen

On Tue, Jan 20, 2009 at 8:45 PM, Ognjen Seslija <oseslija@gmail.com (oseslija@gmail.com)> wrote:

[tomasborrella at gmail...]

Scott, I imagined that it could be an OpenZap problem, but I didn't find an OpenZap mailing list, so I sent the email to FS list. Do you know where can I find more information about OpenZap hardware support and developement status (I have special interest in Loop Start)??

Anthony and Ognjen, I've tried tone detection and thanks to that FS is detecting hung up, but I faced the problem that tone detector answer the call...

That's my dialplan:

<extension name="extension_name">
<condition field="destination_number" expression="^919999999$">
<action application="tone_detect" data="busy 425,0 r +100 hangup 16 4"/>
<action application="bridge" data="sofia/internal/1003%${server-domain-name}, sofia/internal/1004%${server-domain-name}"/>
</condition>
</extension>

When I receive a call from PSTN, tone detection answer the call (the caller hears only one first tone and then hears "nothing" until I pick up the call on softphone).

So, I think that tone detection solution does not resolve my problem... Is there any other possibility to detect hang up without answering the call (using Loop Start signaling) or have we to wait until OpenZap is completely developed?

Thanks in advance.

On Tue, Jan 20, 2009 at 10:43 PM, Ognjen Seslija <oseslija@gmail.com (oseslija@gmail.com)> wrote:

[scott.ellis at novatex...]

I had a similar problem, you can use
<action application="set" data="ringback=${au-ring}"/> (I added an "au" ring definition to my vars.xml file)

To get what you want.

I also had a problem that you get two rings, then an answer then to the system generated ring tone, which was confusing some of our (not to bright) callers.

As we don't use callerID I turned that flag off in the openzap.conf.xml file - I thought that this would do what I wanted (answer the instant the call is detected), but the change in the config file does not make it all the way down to the point where it takes action. At this point I hacked the code to get what I wanted. I have to create a JIRA entry with the details yet.

As far as I understand, this is the right place for OpenZap, as it is a product of the FS project.

Scott

Tomás wrote:

[oseslija at gmail.com]

When call comes in from Openzap, tone_detect app does pre_answer of a call cause it's need media to start detecting tones in the first place. This behaviour is something that I see on calls inside my telco when calling from analogue lines. I don't think this is a big of deal because ringback provided by FS will make caller understand that the call is still in progress. One can make its own ringback to sound exactly the same as telco's.

I don't think that we'll ever make POTS behave like digital protocols do.

Regards,
Ognjen

On Wed, Jan 21, 2009 at 10:36 AM, Scott Ellis <scott.ellis@novatex.com.au (scott.ellis@novatex.com.au)> wrote:

[msc at freeswitch.org]

On Wed, Jan 21, 2009 at 1:36 AM, Scott Ellis <scott.ellis@novatex.com.au> wrote:

[msc at freeswitch.org]

On Wed, Jan 21, 2009 at 1:49 AM, Ognjen Seslija <oseslija@gmail.com> wrote: