[Freeswitch-users] UNIMRCP RECOGNIZE headers

[Neil Burgess]

Hello,

Wondering if anyone can help with a unimrcp question.

We are using the UNIMRCP client in FreeSwitch to communicate with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech Timeouts that are occurring on the server side, however we have been advised that the only way these can be affected is via headers in the MRCP requests. So, we need to be able to set headers such as Speech-Complete-Timeout, Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request. Is there a mechanism in FreeSwitch which we can use to pass these down before, (or as) we issue the "detect_speech" command. We are happy to use a jscript interface, or whatever if such a capability is available.

Many thanks,
Neil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20100107/4c9f7bf4/attachment.html

[Christopher Rienzo]

Neil,

In general, you can set most MRCP params in mod_unimrcp like this:

detect_speech unimrcp {speech-complete-timeout=5000,speech-incomplete-timeout=5000}grammar grammar-name

These params will remain set on the speech handle until freeswitch closes it. For ASR, that's when the call ends.

From: freeswitch-users-bounces at lists.freeswitch.org [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of Neil Burgess
Sent: Thursday, January 07, 2010 10:53 AM
To: freeswitch-users at lists.freeswitch.org
Subject: [Freeswitch-users] UNIMRCP RECOGNIZE headers
Hello,
Wondering if anyone can help with a unimrcp question.
We are using the UNIMRCP client in FreeSwitch to communicate with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech Timeouts that are occurring on the server side, however we have been advised that the only way these can be affected is via headers in the MRCP requests. So, we need to be able to set headers such as Speech-Complete-Timeout, Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request. Is there a mechanism in FreeSwitch which we can use to pass these down before, (or as) we issue the ?detect_speech? command. We are happy to use a jscript interface, or whatever if such a capability is available.
Many thanks,
Neil
_______________________________________________
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

[Arsen Chaloyan]

Hello Neil,


I guess you are using a commercial MRCP server such as Loquendo, at least I'm not aware of UniMRCP server based Loquendo ASR :)
Though someone asked me about such a solution a few months ago.

As of actual request, looking at the code of mod_unimrcp, I'd say such an option exists. See recog_channel_set_params().
Looking through mod_unimrcp wiki examples, I'd say the following should do what you need

switch_ivr_detect_speech(session, "unimrcp", "{recognition-timeout=15000}yesno", "yesno-name", "", ah);
--
Arsen Chaloyan
The author of UniMRCP
http://www.unimrcp.org




________________________________
From: Neil Burgess <neil.burgess at redmatter.com>
To: "freeswitch-users at lists.freeswitch.org" <freeswitch-users at lists.freeswitch.org>
Sent: Thu, January 7, 2010 7:52:52 PM
Subject: [Freeswitch-users] UNIMRCP RECOGNIZE headers


Hello,

Wondering if anyone can help with a unimrcp question.

We are using the UNIMRCP client in FreeSwitch to communicate
with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech
Timeouts that are occurring on the server side, however we have been advised
that the only way these can be affected is via headers in the MRCP
requests. So, we need to be able to set headers such as Speech-Complete-Timeout,
Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request.
Is there a mechanism in FreeSwitch which we can use to pass these down before,
(or as) we issue the ?detect_speech? command. We are happy to use a
jscript interface, or whatever if such a capability is available.

Many thanks,
Neil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20100107/f786ea03/attachment-0001.html

[Neil Burgess]

Hello,

Wondering if anyone can help with a unimrcp question.

We are using the UNIMRCP client in FreeSwitch to communicate with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech Timeouts that are occurring on the server side, however we have been advised that the only way these can be affected is via headers in the MRCP requests. So, we need to be able to set headers such as Speech-Complete-Timeout, Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request. Is there a mechanism in FreeSwitch which we can use to pass these down before, (or as) we issue the "detect_speech" command. We are happy to use a jscript interface, or whatever if such a capability is available.

Many thanks,
Neil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20100107/4c9f7bf4/attachment-0002.html

[Christopher Rienzo]

Neil,

In general, you can set most MRCP params in mod_unimrcp like this:

detect_speech unimrcp {speech-complete-timeout=5000,speech-incomplete-timeout=5000}grammar grammar-name

These params will remain set on the speech handle until freeswitch closes it. For ASR, that's when the call ends.

From: freeswitch-users-bounces at lists.freeswitch.org [mailto:freeswitch-users-bounces at lists.freeswitch.org] On Behalf Of Neil Burgess
Sent: Thursday, January 07, 2010 10:53 AM
To: freeswitch-users at lists.freeswitch.org
Subject: [Freeswitch-users] UNIMRCP RECOGNIZE headers
Hello,
Wondering if anyone can help with a unimrcp question.
We are using the UNIMRCP client in FreeSwitch to communicate with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech Timeouts that are occurring on the server side, however we have been advised that the only way these can be affected is via headers in the MRCP requests. So, we need to be able to set headers such as Speech-Complete-Timeout, Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request. Is there a mechanism in FreeSwitch which we can use to pass these down before, (or as) we issue the ?detect_speech? command. We are happy to use a jscript interface, or whatever if such a capability is available.
Many thanks,
Neil
_______________________________________________
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

[Arsen Chaloyan]

Hello Neil,


I guess you are using a commercial MRCP server such as Loquendo, at least I'm not aware of UniMRCP server based Loquendo ASR :)
Though someone asked me about such a solution a few months ago.

As of actual request, looking at the code of mod_unimrcp, I'd say such an option exists. See recog_channel_set_params().
Looking through mod_unimrcp wiki examples, I'd say the following should do what you need

switch_ivr_detect_speech(session, "unimrcp", "{recognition-timeout=15000}yesno", "yesno-name", "", ah);
--
Arsen Chaloyan
The author of UniMRCP
http://www.unimrcp.org




________________________________
From: Neil Burgess <neil.burgess at redmatter.com>
To: "freeswitch-users at lists.freeswitch.org" <freeswitch-users at lists.freeswitch.org>
Sent: Thu, January 7, 2010 7:52:52 PM
Subject: [Freeswitch-users] UNIMRCP RECOGNIZE headers


Hello,

Wondering if anyone can help with a unimrcp question.

We are using the UNIMRCP client in FreeSwitch to communicate
with a commercial UNIMRCP server (Loquendo). We need to control such items as Speech
Timeouts that are occurring on the server side, however we have been advised
that the only way these can be affected is via headers in the MRCP
requests. So, we need to be able to set headers such as Speech-Complete-Timeout,
Speech-Incomplete-Timeout, Recognition-Timeout, etc in the MRCP RECOGNIZE request.
Is there a mechanism in FreeSwitch which we can use to pass these down before,
(or as) we issue the ?detect_speech? command. We are happy to use a
jscript interface, or whatever if such a capability is available.

Many thanks,
Neil
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freeswitch.org/pipermail/freeswitch-users/attachments/20100107/f786ea03/attachment-0002.html