IBM ANNOUNCES CONTINUOUS SPEECH FOR WINDOWS

NEW YORK, June 29, 1993 . . . IBM Corporation today announced IBM
Continuous Speech Series for Windows, an application development program
that will enable software developers to bring state-of-the-art continuous
speech recognition technology to the tens of millions of PC users
currently operating in the Microsoft Windows computing environment.

"ICSS for Windows represents a significant advance in the rapid evolution
of speech recognition technology," said Elton Sherwin, manager, speech
recognition strategy and market development, IBM Personal Systems. "Speech
is the most natural and effortless means of human communication. Adding a
'speech user interface' to Windows applications provides for a level of
intuitive response that profoundly alters the interaction of man and
machine."

IBM Continuous Speech Series (ICSS) was first introduced in versions for
OS/2 and AIX at Fall Comdex 1992. Since then, IBM has announced a series
of technological breakthroughs in continuous speech, including the use of
inexpensive microphones, low-cost sound cards and the ability to recognize
spoken numbers with a high rate of accuracy. This combination of low cost
and high accuracy will bring speech recognition capabilities to a wide
variety of applications, such as spoken spreadsheets and home bill
paying.

In addition to ICSS for Windows, IBM also announced the immediate
availability of software speech recognition development tools for the OS/2
and AIX environments. These include: ICSS/2 Developers Toolkit and ICSS/2
Runtime System for OS/2 and ICSS/6000 Developers Toolkit and ICSS/6000
Runtime System for AIX.

Speech is ideal for professionals who need to perform repetitive tasks or
input data to computers without looking at a screen or using a keyboard,
such as law enforcement professionals, dentists, brokers, currency
traders, lab technicians and graphic designers. Speech recognition
technology significantly enhances the accessibility of existing data and
applications currently limited to keyboard access alone.

Developed by IBM Gaithersburg, ICSS provides software vendors with an
industry-leading group of tools to create speech-enabled applications.
ICSS tool kits also are available for the OS/2 and AIX operating
environments. Native Windows support for ICSS is scheduled to be available
December, 1993.

Each tool kit includes software that enables:

* Speaker Independence -- Human speech varies significantly from person to
person, and many speech recognition systems require speaker enrollment or
training to be accurate and effective. ICSS does not require training and
can process speech from any speaker who chooses to use it. These
characteristics make it well-suited for kiosk applications.

* Continuous Speech -- Many systems also require that users insert a brief
pause between each word. By contrast, ICSS works through grammar guided
recognition -- it is designed to listen for and process specific sets of
words and phrases -- so the system can effectively and accurately process
English spoken very rapidly. This capability is well-suited for data entry
applications that require high speed control of user interfaces.

* Large Vocabulary -- ICSS operates with an active vocabulary of 1,000
words taken from a base vocabulary of 20,000 U.S. English words.
Developers can easily extend the vocabulary to meet industry-specific
requirements, such as medical or financial terminology.

ICSS may be used either stand-alone or in a client/server configuration and
accommodates a wide variety of microphones.

A developer's program for Windows will make beta level code tools available
to develop automated speech interfaces for Windows 3.1 applications
utilizing OS/2 version 2.0 or 2.1 Native application development tools
Windows mode are scheduled to be available December, 1993.

The developers beta program is available July 23 in three levels, all of
which include software, program code, technical support and education,
users guide and reference and a microphone. The single license program
with a low-cost microphone is priced at $299. The single license program
with a high quality noise canceling microphone is priced at $699. The
multi-license program that includes four copies of the Developers Toolkit
and 20 copies of the Runtime System is priced at $3,995

The ICSS/2 Developers Toolkit and ICSS/6000 Developers Toolkit are priced
at $399 and available immediately. ICSS/2 Runtime System and ICSS/6000
Runtime System are priced at $299 and available immediately. Additional
runtime licenses are available for $249.

In addition to ICSS, IBM's speech recognition product family includes IBM
Speech Server Series (ISSS) which features up to a 32,000-word vocabulary
and takes dictation at a throughput of more than 70 words per minute on a
RISC System/6000 client server platform; VoiceType 2 for DOS applications
such as word processing, databases and spreadsheets; and VoiceType Control
for Windows, for commands and navigation in Windows 3.1.

International Business Machines Corp
IBM United States
1133 Westchester Ave, White Plains, NY 10604

==========================================================
From the America Online -- New Product Information Service
==========================================================
This information was processed from data provided by the
above mentioned company. For additional details, contact 
the company at the address or telephone number indicated.
==========================================================
All submissions for this service should be addressed to:
BAKER ENTERPRISES, 20 Ferro Dr, Sewell, NJ 08080 U.S.A.
Email:  RBakerPC  (AOL),   rbakerpc@aol.com  (Internet)
==========================================================
