Re: [misterhouse-users] MisterHouse and Asterisk

[Robert R. <ro...@fr...>]

----- Original Message ----- 
From: "Gregg Liming" <gr...@li...>
To: <mis...@li...>
Sent: Thursday, February 03, 2005 5:08 PM
Subject: Re: [misterhouse-users] MisterHouse and Asterisk


> Oops... I just realized that I made an assumption about the audio source
> that may not be accurate.  My answer (currently) assumed that the source
> was a mic (wired or wireless) connected to some computer that's running
> sphinx and the voxapl stuff that I'm working on.  An alternative
> possibility that AFAIK does not yet exist is some ulaw/alaw/ilbc/gsm
> encoded source routed by *.   I will eventually get to such a capability
> (unless, hopefully, someone else gets to it first) so that I can control
> MH from my cell phone just be talking to it ;)   But, that's admittedly
> a ways off.
>
I think that Asterisk already has this kind of interface. It's called eagi
and there is example code for connecting with older sphinx (if I remember
right something like eagi-sphinx-test.c), where you can get sound samples
and then probably easiest way would be to send them to Sphinx via TCP
socket....

HTH,

regards,

Rob.



> Quoting Gregg Liming (2/3/05 10:54 AM):
> > Quoting Robert Rozman (2/3/05 10:36 AM):
> >
> >>Gregg,
> >>
> >>I'd like to ask you another thing. If I remember right, you're also
working
> >>on sphinx integration and MH. In what way would it be easiest to connect
> >>Asterisk and Sphinx, so to recognize some common words, like saying name
of
> >>desired extension like "talk to Mary" ?
> >>
> >
> >
> > Wow--I hadn't thought of that, but definitely doable.  The current
> > implementation will allow you to map any sphinx grammar that you define
> > to any xAP (and soon xPL) schema that you might think up (or that
> > exists) on a per grammar basis (w/ grammar switching options).  I'd
> > first take a look at the xAP BSC and also CID-related schemas as a point
> > of departure; but, a new one may be required for this application.  The
> > only "down side" is that the person's name would need to be defined as
> > part of the JSGF grammar so as to maximize VR accuracy.  Also, for
> > uncommon names, the phoenetic definition may be necessary as a new
> > dictionary entry.  However, IMO, all of this is the simple part.
> >
> > The part that doesn't yet exist (AFAIK) is an xAP listener that can
> > control *.  I rather liked Jason's original approach to integrating w/
> > the * manager rather than AGI.  This is as I recall the same approach
> > that was used by that flash project that you've mentioned before.  So, I
> > think that the next issue to look at is the API-specifics w/i the
> > manager that can perform the control that you want.  If you (or others)
> > are willing to look at the * end of things, then I could assist w/ the
> > xAP/manager integration and definitely w/ providing you with an initial
> > grammar "starter" kit--essentially, the grammar, dictionary and schema
> > mapping files that would be needed.
> >
> > Just wondering out load... does this mean that the manager would attempt
> > to dial "Mary" and then connect up to some local extension that would be
> > defined either via the grammar or a prior in *?  I'm just wondering
> > about the logistical end of setting up the call and handing it off to
> > the appropriate phone/phones--or, were you thinking of setting up an
> > agent queue?  I'm also thinking that maybe the xAP/* listener ought to
> > send off a xAP TTS and/or OSD message once the dial is attempted to
> > provide feedback that it understands what you've requested and is
> > attempting to service the request.
> >
> > If you're interested, this probably ought to be an entirely different
> > thread as there's a lot to iron out.
> >
> > Gregg
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
> > Tool for open source databases. Create drag-&-drop reports. Save time
> > by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
> > Download a FREE copy at http://www.intelliview.com/go/osdn_nl
> > ________________________________________________________
> > To unsubscribe from this list, go to:
http://sourceforge.net/mail/?group_id=1365
> >
> >
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting
> Tool for open source databases. Create drag-&-drop reports. Save time
> by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc.
> Download a FREE copy at http://www.intelliview.com/go/osdn_nl
> ________________________________________________________
> To unsubscribe from this list, go to:
http://sourceforge.net/mail/?group_id=1365
>