Re: DRAGON-10 VOICE RECOGNITION PHYSICS
George Hammond
Xl.tags.giganews.com!border1.nntp.dca.giganews.com!nntp.giganews.com!local2.nntp.dca.giganews.com!news.giganews.com.POSTED!not-for-mail
NNTP-Posting-Date: Sat, 12 Sep 2009 17:18:47 -0500
From: George Hammond <Nowh...@notspam.com>
Newsgroups:
sci.physics.relativity,sci.physics,alt.sci.physics.new-theories,alt.sci.physics,rec.org.mensa
Subject: Re: DRAGON-10 VOICE RECOGNITION PHYSICS
Date: Sat, 12 Sep 2009 18:18:50 -0400
Message-ID: <r36oa5tdp7nlargpn...@4ax.com>
References: <1s9ma5dscr51rlpi6...@4ax.com>
<3c7dad72-f108-475a...@o9g2000yqj.googlegroups.com>
X-Newsreader: Forte Free Agent 3.1/32.783
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Lines: 75
X-Usenet-Provider: http://www.giganews.com
X-Trace:
sv3-lusHUYn1jDlPL87hGOSbjVsIaYAqmSmS+nG+rWHO20G6UnvNNZMrTxq48S6kT9jkchsE5b7e+m96Z8Q!tqJUgPWise+r6ImO6j9ds2I1QnQut669z8KnnT8tnOj13duknO96IlE0L/dqMW9oOyMSSw==
X-Complaints-To: ab...@giganews.com
X-DMCA-Notifications: http://www.giganews.com/info/dmca.html
X-Abuse-and-DMCA-Info: Please be sure to forward a copy of
ALL headers
X-Abuse-and-DMCA-Info: Otherwise we will be unable to
process your complaint properly
X-Postfilter: 1.3.39
Bytes: 4437
Xref: number.nntp.dca.giganews.com rec.org.mensa:477733
alt.sci.physics:184577 alt.sci.physics.new-theories:253051
sci.physics:1563527 sci.physics.relativity:927807
On Sat, 12 Sep 2009 03:41:39 -0700 (PDT), Harbinger
<jhngui...@googlemail.com> wrote:
>On 12 Sep, 06:28, George Hammond <Nowhe...@notspam.com> wrote:
>> � �I am dictating this post using Dragon Naturally
>> Speaking-10. �I am planning to use the speech recognition
>> program to dictate a 300 page book.>
>
>
>Dear George,
>
>Good to hear from you. From time to time, I have wondered about speech
>recognition, though left the matter in abeyance, simply because it
>seemed to be going nowhere boastfully and at near light speed!! If it
>is as good as you say, then we all need to take another look! I have
>always thought that the barrier is the Cocktail Party Effect v speech
>recognition of robotic precision!?
>
[Hammond]
Well, the first thing you notice is the sheer size of
this Dragon-10 program, 2 GB of hard drive space! I think
this is considerably larger than previous versions of
Dragon. I suspect that the recent advances in search engine
programming, Google for instance, has produced sophisticated
ways of analyzing the syntax, grammar, content, etc. of
sentences. Speech recognition after all in the human brain
is not merely a dictionary lookup process of acoustic input,
it is a sophisticated analytical process involving language
structure. I think the addition of this new high-technology
to Dragon is what has made the difference in the power of
their new version 10.
I know one thing, I'll never go back to manual typing now
that I've seen this.
>
>
>
>Anyway, albeit piss-poor, wreck.org.mensa ain't that bad. We ought to
>look after one another better! The bright shouting one another down is
>not my idea of good. And, yes, good is an indefinable!
>
>
[ Hammond]
Another thing I've discovered, is that strong writers are
strong speakers. Stephen Jay Gould, Richard Dawkins, Ernest
Hemingway, and even Sigmund Freud were animated and
energetic speakers who could talk nonstop for hours.
Such people are not highly analytical, but are focused
more on perception and seeing than deep logical intricacy.
What I have discovered is that one's speaking style or
even dictating style is considerably different from one's
writing or typing style. if enough concentration is used to
formulate the sentences before they are spoken one can
actually achieve a very commanding presence in a written
style. This is something that I working on at the moment
since I plan to dictate a 300 page book entitled:
THE
SCIENTIFIC
PROOF OF GOD
Anyway, I have just dictated this response to you
directly into Forte Agent in less than two minutes. And so
far it's only made one or two mistakes. I find that utterly
amazing if not flat out MAGIC!
========================================
GEORGE HAMMOND'S PROOF OF GOD WEBSITE
Primary site
http://webspace.webring.com/people/eg/george_hammond
Mirror sites
http://geocities.com/scientific_proof_of_god
http://proof-of-god.freewebsitehosting.com
HAMMOND FOLK SONG by Casey Bennetto
http://interrobang.jwgh.org/songs/hammond.mp3
=======================================
cocktail
I understand that Windows-7 (is that Vista?) has a
speech recognition program built right into it and it's
actually listed on the Start-Menu!
How does this Microsoft speech recognition program stack
up against Dragon Naturally Speaking-10.
Which is better?
George Hammond
<ftl...@hotmail.com> wrote:
>
>>
>> PS; Dragon-10 is on sale nationally at half price. You can
>> order a copy from Amazon.com for $44, that's where I got
>> this one. It's a 2 GB program on a CD. It loaded into my
>> 160 gig hard drive no problem.
>>
>> PPS: I think their slashing the price has something to do
>> with the Vist program, they probably think Bill Gates will
>> drive them out of business? Just a guess, because this
>> program for $44 is the biggest software bargain I've ever
>> seen!
>
>They probably lowered the price because version 11 is going to be released.
>Standard tactic...
>
>
>
[Hammond]
Na... I doubt it, because they never slashed prices on
versions 5, 6,7,8 or 9 when they released 10. Besides there
is no 11 in the works.
Fact is Windows-7 (I think it's called Vista) has a
speech recognition program built right into it and its even
listed right on the Start Menu according to what I hear. Do
you know how many copies of Windows-7 there are going to be
running in a few years..... christ....Bill Gates could put
all the independent voice recognition companies out of
business... and Dragon must be putting up a fight to get
their product (which is better than Microsft's) out there
and recognized! Dragon could wind up being another
Microsoft fatality like WordPerfect.... remember them?
Have you seen a Windows-7 computer? I saw a Microsoft
demo of their speech recognition on national t.v. on the 5
o'clock news... and would you believe they couldn't get it
to work and it kept printing out lines of gobbledygook!!
Mark Edwards
> Anyway, I have just dictated this response to you directly into Forte
> And so far it's only made one or two mistakes.
IF[HAMMOND]PFH.
Mark Edwards
--
Proof of Sanity Forged Upon Request
George Hammond
<Mark-E...@comcast.net> wrote:
>No cluons were harmed when George Hammond wrote:
>> Anyway, I have just dictated this response to you directly into Forte
>> Agent in less than two minutes, and so far it's only made one or two mistakes.
>
>
>with only one hour spent proofreading
>Mark Edwards
>
>
[Hammond]
Na, the thing has a built-in spellchecker and auto word
correction and you simply read it on the screen as you speak
it so there is no need for proofreading.
The nice thing about it is to you can sit here and rap
and rap and rap saying absolutely nothing and drown out
hecklers with thousands of words that cost you absolutely no
effort to post while they are still pounding their Stone Age
keyboards and beating their fingers raw. There is nothing
like superior firepower.
Right now I'm in a very exclusive club using this thing
and I can really kick kook ass all day long. But Windows 7
is going to be released in a few months and that is going to
have a word processing speech program built into it right on
the start menu. Once the loudmouth spammers get a hold of
that USENET is going to be drowned in Heckler spam. Thank
God I'll no longer have to listen to it and watch the final
downfall of USENET since I'm busy dictating a 300 page book
into this thing and am now spending my time talking to the
reading public who can't talk back and have to pay for the
privilege of listening to me.
George Hammond
<jpol...@cfl.rr.com> wrote:
>On Sat, 12 Sep 2009 01:28:59 -0400, George Hammond
><Nowh...@notspam.com> wrote:
>
>> I am dictating this post using Dragon Naturally
>>Speaking-10. I am planning to use the speech recognition
>>program to dictate a 300 page book.
>>
>>THE
>>SCIENTIFIC
>>PROOF OF GOD
>>
>> Are there any physicists on these newsgroups who have
>>taken a serious look at commercially available speech
>>recognition programs?
>>
>
>
>
>
>[John Polasek]
>George I have been using NatSpeak 9.5 for a couple of months and I'm
>extremely pleased, but it will not tolerate mumbling. This program is
>a surprisingly efficient verbosity enhancer.
>
>I'm surprised you could load version 10 with a CD because I had v10
>and had to attach a DVD drive, and even then it refused to load
>indicating I didn't have enough computer (2 gig Sempro 1.1 gig
>memory). So I returned it and got 9.5 on a CD and it loaded slickly. I
>have not seen a cogent discussion that outlines the computer
>requirements, but I suspect a high-speed dual channel memory would
>help since the program is basically scanning down a list of phonemes.
>
>I find the program needs a lot of help. One tremendous aid is the use
>of Microsoft Word's AutoCorrect, which lets you enter a bunch of
>aliases in one column and what you would really like to see typed in
>the other column. In addition, my macro keyboard has a join function
>(to fix when you pause and it starts a new sentence) and a
>recapitalization function.
>They provide essentially no support and I haven't found a useful blog
>that didn't have ancient entries like 2007.
>I would estimate that this message took 17 (minimum) corrections.
>Maybe we could trade tips. This program needs a "morning after"
>capitalization function.
>.
>John Polasek
>
>
>
[Hammond]
My experience is a good luck story apparently. I got
Dragon-10 on sale for only $44 from Amazon.com after they
slashed their prices. I'm running it on a Compaq Computer
with an AMD 3400+ 2 gig processor. I have a 160 gig hard
drive so space is no problem. It has a DVD/CD player built
in. Believe it or not I only have 512 megs of RAM! Yet the
thing is running perfectly.
It runs best if you shut everything down and just type
into Dragon pad. But right now I am typing directly into
forte agent and is working excellent.
I haven't even read the instruction manual yet believe it
or not. I WAS LUCKY, I JUST SLAPPED THE CD IN THE DRIVE
loaded it and started talking and everything worked without
a single glitch. My computer is only four years old, that's
probably why.
I bought it mainly because I plan to dictate a 300 page
book entitled:
THE
SCIENTIFIC
PROOF OF GOD
I have measured the speed with a word counter and stopwatch
and I have had it up to as high as 120 words per minute and
that's only because that's about as fast as I can talk!
However, when mentally composing and dictating my book I
find that I usually don't go much faster than 50 words per
minute, so I don't think I'm going to have any problem. I
looked into shoving another couple of gigs of ram into this
computer and maybe I will but the fact is I can live without
it. This thing is working fine with only 512 megs. I saw
somewhere someone make a comment that Dragon rarely uses
more than 300 megs. It's when you open up other
applications and start multitasking that you run out of
memory.
I have started loading in a special technical vocabulary
(physics) and I am now studying the voice command structure
which I haven't been using at all.
Give me another couple of weeks to master this thing, and
yes I'd like to trade information with you about Dragon in
the future. I hang out on s.p.r. but have crossposted this
to comp.speech.users.
Are there any USENET groups that discuss Dragon speech
recognition besides comp.speech.users?
Mark Edwards
> There is nothing like superior firepower.
Intellectually, this is true. Sorry to hear you came unarmed...
<SNIP HECKLER CRAP>
RVince
also use it to read in lists of numbers, such as 3,127,001.06 and it
does all of this perfectly.
My primary use of it is as a "front end" to a program I have been
working on for years that acts as the muscle to speech recognition,
i.e. parsing the speech for meaning.....
"Jack is a mouse, Sharon is a bird. Mice are little. Is he small?"
To which the response would be, "Yes, Jack is small because Jack is a
mouse and mice are little."
Incidentally, the voice synthesis I use (which is the best anaolog I
could find to Dragon for SR) is the open sourced MARY project.
Together, with Dragon, and ideal combination.
-RVince
Otto Bahn
--oTTo--