From: Mike R. Manzano (alephx01_at_mac.com)
Date: Wed Feb 16 2005 - 15:53:01 PST
There are two major applications of voice recognition:
* voice dictation -- using voice recognition to transcribe text
* voice command -- using voice recognition to issue imperative
instructions that the computer will act on
The latter type usually also involves natural language understanding,
which is the transformation of natural language, whether written or
spoken, into a form a computer can understand.
I can see voice dictation being useful for people who have a disability
that prevents them from writing. However, because the input vector for
voice dictation is usually much noisier than input vector for HWR
(think, microphone in a noisy room vs. touch-screen in a noisy room),
HWR will usually be the better of the two. Besides, nobody wants to
hear you recite your emails as you write them. There has been some
research into subvocalization, but that's another subject altogether.
I see voice command as being the more promising of the two as it
contributes to a multimodal interface. By multimodal, I don't mean
voice commands that do the same thing as your mouse or keyboard would
do (i.e., there's no real point to saying "File Menu. Open Menu." to a
computer versus just using the mouse to do it). What I mean by
multimodal is the use of voice command to do things that are not
immediately already taken care of by the mouse or keyboard. For
example, if you have a Mac, turn on Speech, click on a document in the
Finder, and say "send this to <contact>", and it will look up
<contact>'s email address and send them the document. This is a good
use of voice because the alternative would be to drag the icon to your
mail app, look up the address in your address book, and enter it into
the To: field, and hit "Send". For the Newton, I can imagine such
commands as "show email address for John", or "call Mary". Even more
useful, being able to say "delete this paragraph" or "italicize the
previous sentence" while writing a note, like you can do with IBM
ViaVoice or Dragon.
Regards,
Mike
On Feb 16, 2005, at 1:44 PM, Adriano wrote:
> What do you think about voice recognition?
>
> If i remember well, it's enabled in the NewtonOS (?)
>
> I have seen a lot of commercial software for OS 9 and OS X for enabling
> voice recognition for specific software (iTunes, iPhoto, etc etc etc).
>
> I think that the MessagePad could become one of the best platform for
> people with handicaps,
> just using its own microphone is not the best, but D.Padilla's did the
> most enanchements on helping the voice/sounds recognition in the
> MessagePad.
>
> Adriano
>
> Il giorno 16/feb/05, alle 22:18, Mike R. Manzano ha scritto:
>
> Even with the 2x00 handwriting recognition is slower and less
> accurate=20=
>
> than typing, especially for touch typists.
-- This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries Official Newton FAQ: http://www.chuma.org/newton/faq/ WikiWikiNewt for all kinds of articles: http://tools.unna.org/wikiwikinewt/
This archive was generated by hypermail 2.1.5 : Wed Feb 16 2005 - 18:00:01 PST