The UK Home Automation Archive

Archive Home
Group Home
Search Archive


Advanced Search

The UKHA-ARCHIVE IS CEASING OPERATIONS 31 DEC 2024


[Message Prev][Message Next][Thread Prev][Thread Next][Message Index][Thread Index]

Re: vendor ID reservation


  • Subject: Re: vendor ID reservation
  • From: Gregg Liming <gregg@xxxxxxxxxxxxxxxxxxxx>
  • Date: Mon, 17 Jan 2005 22:27:15 -0500
  • References: <D076CE2AFAB6334395AB4466654BC4DF0A936E@turing.Wintermute.local> <41C02750.4090604@limingconsulting.com> <006001c4e29f$827eec10$040612ac@corp.frontier.co.uk> <41C02FBF.7040101@limingconsulting.com> <26b8eee105011716181f866d8a@mail.gmail.com>


Quoting Tom Van den Panhuyzen (1/17/05 7:18 PM):
> On Wed, 15 Dec 2004 07:36:15 -0500, Gregg Liming
> <gregg@xxxxxxx> wrote:

>>I can't speak for the others... but, I'm working on
"wrapping" the
>>Sphinx 4 voice recognition engine

>>I hope to be using this live in my house by Christmas; so, sharing
with
>>everyone else by that time frame is likely.

> Ha! The virtues of gmail.  Easily search archives. :-)
> I wonder if Gregg made any progress ?

Yes--but, not as much as I'd hoped.

> I really would like to put a big seat in the middle of the living room
> one day and from there say "Computer! Take us home! Warp speed
9!"
> All the lights go out and of course the computer answers with that
> lovely female voice.

Well, I'm doing something like that now--only w/ xAP and misterhouse
(the starting targets).  What I did get accomplished before the holidays
was a fairly major rewrite that allows schema mapping between schema
attributes from xAP or xPL (xPL is not yet tested) and  attributes
associated w/ utterances w/i a grammar.  This is hard to explain w/o
actually seeing a real example, but the outcome allows for different xAP
or xPL (soon) messages to be sent on a per utterance basis.  In general,
you wouldn't be switching constantly "on the fly", but you would
want a
change when changing grammars--which was the other major design rewrite
(specifically, dynamic grammar switching).  As an example, the utterance
"Computer--take us home" (no perceivable break/pause in phrase)
might
switch to the "home grammar" and the utterance "Warp speed
9!" is
defined w/i the home grammar w/ mappable attributes that produce a
command message.  The distinction here is that perhaps "Warp speed
9" is
also an utterance that could appear in a different grammar--like
"bartending/ cooking"--and resulting in a command to your xPL
controlled
blender to mix your favorite libation.

I hope to begin testing in earnest the pre-holiday work this week to
confirm that the xPL grammar mapping works.  My personal interest would
be xPL TTS and/or xPL audio.basic (for initial xPLRioNet control).
However, I'm guessing that the "control" analog that I currently
use for
misterhouse is the new (?) xPLHal ability to fire a determinator (right
term?) based on some xPL message.  Can anyone point me to the xPL schema
that supports this capability? I'm guessing that is where most
(non-misterhouse) xPL users would find the most interest.  As an
example, the scenario that Tom mentions above would actually require
multiple xPL commands to fire (TTS and the lights out (X10?)) .  The app
that I've written only maps a single xAP or xPL message to a given
utterance; so, chained or complex "action" logic is considered
not w/i
scope and must be implemented by the xPL message recipient (e.g.,
xPLHal).  If anyone could suggest a good test case--or, better, yet
determinator script that I might use for testing, then I'd start w/ that.

FWIW: I currently use a "musician grade" wireless (VHF) mic that
looks a
bit like what you'd expect for Karaoke (evoking smirks from observers ;)
).  This was deliberate because the mic can be switched on or off
(off-means that the app's CPU use goes to almost zero) and is quite
mobile.  Remarkably, I've not yet had sphinx recognize the TTS
prompts/acknowledge messages (thereby triggering secondary actions based
on supporting TTS).  It's probably worth pointing out that microphone
quality and the lack of excess background "noise" is key to
getting
reliable recognition.

I plan to follow w/ more useful info on how to use voxapl once it's at a
point that is ready "for the masses" (i.e., tested and at least,
somewhat documented).   If anyone is more anxious to take a look, then
you can grab source from cvs at http://sourceforge.net/projects/voxapl.
If you have specific interests/features in mind, the please do let me
know.  I doubt I'll be soliciting a vote--though ;)

Gregg


xPL Links: http://www.xplproject.org.uk http://www.xplhal.com http://www.xpl.myby.co.uk
To Post a Message: ukha_xpl@xxxxxxx
To Subscribe:  ukha_xpl-subscribe@xxxxxxx
To Unsubscribe:  ukha_xpl-unsubscribe@xxxxxxx

xPL Main Index | xPL Thread Index | xPL Home | Archives Home

Comments to the Webmaster are always welcomed, please use this contact form . Note that as this site is a mailing list archive, the Webmaster has no control over the contents of the messages. Comments about message content should be directed to the relevant mailing list.