Re: Homeseer and voice

["Oliver Pell" <odp@xxxxxxx>]

I've done quite a bit of experimentation with Homeseer and voice
recognition
and I'm not at all convinced you're going to get any results worth writing
home about after spending £50 :-(. Speakers really don't matter - if you're
only using them for TTS anyway - but the quality of the microphone does. I
spent a couple of weeks trying a headset plugged into my computer but (as
expected) it was only any use if I was wearing the headset and the sound
quality was lousy. In the end I shelled out for an Audiotec Pro42 mic from
Stage Electrics and got a phantom power source from LetsAutomate - voice
recognition does work all around the room now. The sound volume is really
low - but obviously high enough for the computer to understand. I was
getting 80-90% command recognition, I would say.

My current plans are probably to get one of LetsAutomate's mixers so I can
connect more than one mic (and hopefully that'll boost the volume as well).

Have you tried the voice recognition so far? Homeseer comes with the MS
voice recognition engine which only seems to really work well (as far as VR
ever works well) with American accents. I would recommend getting a hold of
a commercial VR program, training it, and then using that engine - I'm
using
the Dragon NaturallySpeaking engine and the difference between it and MS is
immense. There is then the slight problem that other people can't really
use
it without you putting in some fancy VBScript code to switch speakers but
this is a minor problem compared to the increase in recognition quality
(for
me, anyway).

Hope some of this helps,

Oliver.

Message: 18
Date: Thu, 26 Apr 2001 10:00:42 +0100
From: "Gareth Cook/STA/Lotus" <gcook@xxxxxxx>
Subject: HomeSeer and Voice

ok - I've installed HomeSeer. I have a server in my loft, and I'd like to
use voice recognition and text to speech.

So - where do I start ? Any recommendations for microphone and speakers,
and placement, and shielded cable types ?
<snip>