When I Talk, My Computer Types!

birthdaytestΤεχνίτη Νοημοσύνη και Ρομποτική

17 Νοε 2013 (πριν από 3 χρόνια και 4 μήνες)

48 εμφανίσεις

When I Talk,

My Computer Types!

By Karl Barksdale

After more than three decades of speculation about when speech recognition
software would be ready for prime time, Continuous Speech Recognition (CSR)
programs are now gaining popularity. With increased p
rocessor speeds,
declining memory costs, and increased storage capacity, the personal computer
can now “type” (key) in response to the spoken word.

The voice revolution began in the 1960s at IBM’s famed Thomas J. Watson
Research Center in Hawthorne, New
York. IBM has long believed that the next
jump in computer productivity would be caused by a voice
interface revolution.
To this end, they committed three decades of research to study voice
recognition. And they were not alone in their belief in the potent
ial of voice
typing. While IBM was introducing its CSR software called
, other
companies were introducing CSR systems of their own.

Dragon Systems impressed customers and analysts alike with

Lernout & Hauspie excited the crowd


Philips shone the spotlight on its

Microsoft introduced
Microsoft Speech Recognition

Dozens of smaller companies began to compete in the vigorous CSR

The voice recognition software now available ha
s progressed to the point that,
when you talk, your computer can key your words accurately. CSR is efficient
for students, consumers, and workers in many professions.

The Dream Versus the Reality

Star Trek
made talking to computers look so easy.

are never as easy as they appear in the movies. The truth is, it takes time
and work to learn the basics of a CSR program. After the basics have been
conquered, it takes additional time for dedicated practice to train your computer
to understand your uniqu
e way of speaking.

The results can be astounding. A typical student who spends adequate time
training the software can expect to voice
key between 110 and 160 words per
minute (wpm) with 90 to 95 percent accuracy. Voice
keying of over 130 words
per minute

with accuracy exceeding 96 percent is commonplace after training.

Health and Safety Issues

Beyond the obvious input efficiencies for data entry, there are other reasons why
businesses and schools are accepting voice. There is concern about the
e of repetitive stress injuries (RSI) and carpal tunnel syndrome in the
When I Talk, My Computer Types!



workplace. The Occupational Safety and Health Administration (OSHA) calls
these injuries
related musculoskeletal disorders
, or WMSDs.

It is interesting to note that the sharp rise

in WMSDs corresponds with the
infusion of personal computers into the workplace starting in the mid
While most WMSDs still occur in the manufacturing sector, office workers,
using computer keyboards and the mouse, now suffer a significantly high
rtion of WMSDs. Speech recognition technology represents a very important
weapon in the battle against WMSDs. While it is possible to take preventative
measures against WMSDs (e.g., frequent breaks, hand and finger exercises, wise
choices in office furnitu
re), voice technology does take stress off the hands.

Voice dictation can help office workers who depend on keyboarding for their
livelihoods. CSR systems can dramatically reduce the number of repetitive
keystrokes these workers must make each day. Does t
his mean that some office
workers can quit keying and use voice instead? For many suffering with
WMSDs, the answer is an emphatic YES! For many carpal tunnel syndrome and
RSI sufferers, voice is a highly effective alternative to keyboarding.

Blending Keyb
oarding with Voice Input

Even on
Star Trek
, Mr. Spock and Commander Data key on their consoles from
time to time.

Most of us will continue to key part of the time and use our CSR software the
rest of the time. It is therefore essential that keyboarding
instruction continue to
emphasize correct keying techniques and proper computer
office ergonomics. In
some situations it will be difficult, if not impossible, to use a CSR program; for
instance, taking notes on your computer during a lecture or a sales con
Also, speaking all day to a computer can cause problems with your vocal cords.
If you are going to use voice dictation software, you had better keep your water
bottle handy and give your voice a rest by keying with your hands every once in
a while

However, if you use a CSR program for even 50 percent of the time you spend
in front of a computer, your chances of suffering severe WMSDs as you get
older will diminish. So, if for no other reason than to avoid future long
term pain
and suffering, a C
SR program is well worth learning. Millions of people are
already using voice recognition technology

at school and on the job.

Now owned by Lernout & Hauspie (L&H)

Hardware and Software Requirements

Based Hardware

Pentium II 266 MHz or

2 333 MHz or higher processor

64 MB of RAM

200 MB of hard drive space

10 additional MB of hard drive space for each new student on that
specific computer

Creative Labs Sound Blaster 16 or other approved sound card

When I Talk, My Computer Types!



Based Har

Pentium III 450 MHz equivalent or higher processor

128 MB of RAM

300 MB of hard drive space

10 additional MB of hard drive space for each new student on that
specific computer

Creative Labs Sound Blaster 16 or other approved sound card

Minimum M
Based Hardware

350 MHz G3 processor or higher

64 MB of RAM (128 MB of RAM recommended)

300 MB of hard drive space

10 additional MB of hard drive space for each new student on that
specific computer

This article has been updated since it appea
red in
Voice Technology Basics
, ©
2000, South
Western Educational Publishing. We would like to thank the
author, a recognized Continuous Speech Recognition authority and enthusiast.


Reprinted from
Century 21 Computer Appli
cations & Keyboarding, 7