Re: Scripting languages

Joerg <notthisjoergsch removethispacbell.net> wrote:

John Doe wrote:
Joerg <notthisjoergsch removethispacbell.net> wrote:


John Doe wrote:


Joerg <notthisjoergsch removethispacbell.net> wrote:


But I don't want to condition my voice so that some machine
understands it.


I couldn't care less whether a machine understands what I say, and
that's not the goal. Maybe that just seems like the goal.


Then how would a machine that didn't understand your command
execute the macro that you wanted?


The distinction should be made because "recognition" is far from
"understanding". It does the recognition thing and then matches
that with the most likely possibility from a list of words in its
vocabulary. Word popularity is a heavy factor.

Whatever you call it, the objective is not to get a machine to
understand what you say. A singer conditions his voice with the
goal of wooing an audience, not to make the sound waves move
through the air in a particular way.


However, no matter how you turn it in the end the machine has to
recognize (or whatever you'd like to call it) your command and then
execute the correct macro. If it works like some voice recognition
"customer service" outfits it's 50% hit or miss.

Get real. For the vast majority of users, speech-recognition automated
phone service must be much greater than 50% hit or miss. You're dealing
with the wrong company.

Doctors and other medical professionals use speech recognition.

One thing I use speech recognition for is online gaming. Of course
that's not critical, but it gets very intense. My speech activated
macroing is absolutely positively extremely accurate, and it uses a
much larger vocabulary than a typical speech driven phone service
implementation.

Ham radio operators (I am one)

Not according to your other reply.

know how tricky that can be even for
a human when dealing with noise plus a heavy accent. So they, the
military guys

Have you ever been in the military?

and lots of others use special words. An example: They
say "niner" instead of "nine", to avoid a mix up with "five".

There's nothing unusual about using a lesser technology in a more
critical environment. Using speech recognition, you learn to fluently
revert back to doing things by hand and then back to using the
technology.

It is quite tough to make voice recognition work reliably without at
least some modification of speech.

That depends a lot on the vocabulary. If you are getting 50% accuracy
from a decent implementation of speech driven phone service that has a
few word vocabulary, Yes you should forget about using speech
recognition on a computer for dictation or macro activation.

Also, simply dismissing accents ain't going to fly anymore,
considering that more than a third of Californians have some kind of
accent. Heck, even our governor does. IOW, machine recognition has
to be able to deal with that or it's not working right.

Oh whail. If you're rich, fund the research to make it work better. If
not, either cope with the fact you cannot use it or train your voice
so that you can.














--
Regards, Joerg

http://www.analogconsultants.com

Path: newsdbm02.news.prodigy.net!newsdst02.news.prodigy.net!
newscon02.news.prodigy.net!prodigy.net!newsdst01.news.prodigy.net!
prodigy.com!postmaster.news.prodigy.com!
newssvr11.news.prodigy.net.POSTED!78df62d6!not-for-mail
From: Joerg <notthisjoergsch removethispacbell.net>
User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; rv:1.6)
Gecko/20040113
X-Accept-Language: en-us, en
MIME-Version: 1.0
Newsgroups: sci.electronics.design
Subject: Re: Scripting languages
References: <13cufcj2i2tq7c9 corp.supernews.com>
<h7huc350695eulqgdihfdi57v2eeu2sef5 4ax.com> <46CF549F.9090104
SpamMeSenseless.pergamos.net> <13cuv22h4v766e0 corp.supernews.com>
<1188225692.806005.172290 19g2000hsx.googlegroups.com> <XdEAi.56$3Y1.10
newssvr17.news.prodigy.net> <YlEAi.50$4J3.8 newssvr22.news.prodigy.net>
<fxEAi.63$3Y1.58 newssvr17.news.prodigy.net> <AJEAi.21585$eY.13393
newssvr13.news.prodigy.net> <w9HAi.434$FO2.325
newssvr14.news.prodigy.net> <AqHAi.438$FO2.244
newssvr14.news.prodigy.net> <eOHAi.65$4J3.5 newssvr22.news.prodigy.net>
In-Reply-To: <eOHAi.65$4J3.5 newssvr22.news.prodigy.net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 54
Message-ID: <nrIAi.30719$RX.29900 newssvr11.news.prodigy.net>
NNTP-Posting-Host: 75.26.178.215
X-Complaints-To: abuse prodigy.net
X-Trace: newssvr11.news.prodigy.net 1188254547 ST000 75.26.178.215
(Mon, 27 Aug 2007 18:42:27 EDT)
NNTP-Posting-Date: Mon, 27 Aug 2007 18:42:27 EDT
Organization: SBC http://yahoo.sbc.com
X-UserInfo1: TSU[ ION_ZUYB_TY]ZKTOFTBTR\B GXLN GZ_GYO^ZUDUWYAKVUOPCW[ML
\JXUCKVFDYZKBMSFX^OMSAFNTINTDDMVW[X\THOPXZRVOCJTUTPC\_JSBVX
\KAOTBAJBVMZTYAKMNLDI_MFDSSOLXINH__FS^\WQGHGI^C E[A_CF\AQLDQ
\BTMPLDFNVUQ_VM
Date: Mon, 27 Aug 2007 22:42:27 GMT
Xref: prodigy.net sci.electronics.design:835519
X-Received-Date: Mon, 27 Aug 2007 18:43:01 EDT
(newsdbm02.news.prodigy.net)



.

Relevant Pages

  • [opensuse] Re: Speech-to-Text; Qs and VV backstory...
    ... I've found a couple of proprietary apps but you have to use Voice mail as an input. ... recognition -- but even then, after training, it still was pretty ... before they recovered and had a continuous speech product. ... For application integration, though, IBM only added full ...
    (SuSE)
  • Re: Scripting languages
    ... The distinction should be made because "recognition" is far from "understanding". ... phone service must be much greater than 50% hit or miss. ... much larger vocabulary than a typical speech driven phone service ... There's nothing unusual about using a lesser technology in a more ...
    (sci.electronics.design)
  • Re: Speech Recognition Toolkits - Requirements
    ... Despite the variety of ASR and SAPI systems around, ... mess up the operation of speech syntheses, so the ASR must probably have ... As you can see, integration of recognition and synthesis, while being ... These tools help the developer build a language-specific ASR Engine around a set of mathematical models, providing required APIs. ...
    (comp.speech.users)
  • Re: NewLine bug of Windows Speech Recognition revisited
    ... especially where spelling is concerned and that is with any app! ... "Queen's English" (going on how long it has taken you to learn speech ... In reading into certain posts, most people in here including myself think ... so long to learn how to use speech recognition and more importantly SPELLING ...
    (microsoft.public.windows.vista.general)
  • Windows SAPI5 Speech Recognition
    ... text to speech, but I'm having trouble gettng the speech recognition part to ... ;; From the SAPI5 manual: ... The following "Recognition Event" information comes from the ... StreamNumber As Long, ...
    (comp.lang.lisp)