interesting article in the FT June 2 about speech ,copied from Yahoo
board.
Article from Financial Times june, 2nd
by: Bambari
8197 of 8201
LERNOUT&HAUSPIE = world leader !!!
globalarchive.ft.com
----------------------------------------------------------------
SURVEY - FT INFORMATION TECHNOLOGY REVIEW: The power of
speech in the digital age: VOICE RECOGNITION by Paul Taylor: The
technology has spent years in the laboratory, but is now becoming mature enough
to enter the commercial mainstream
81% match; Financial Times ; 02-Jun-1999 10:05:17 am ; 1311 words
----------------------------------------------------------------
The first electronic speech synthesiser dubbed "Voder" (voice encoder) was
developed in 1936 by researchers at AT&T's Bell Labs and speech chips were
popularised by Texas Instruments in their "Speak and Spell" toys of the 1970s.
But despite Hal 9000, the talking computer in 2001: A Space Odyssey, Stanley
Kubrick's classic film, voice recognition has proved a much tougher technology to
crack. Over the past few years, however, speech recognition, voice synthesis and
other voice technologies have made dramatic progress. After years of laboratory
development, voice technology is now a hot topic, poised to enter the commercial
mainstream.
part 2 - FT article
by: Bambari
8198 of 8201
"1999 will be a breakthrough year for speech technology," says Gaston
Bastiaens, president and chief executive of Lernout & Hauspie, the
Belgian-based speech technology market leader. "During the past year, the
market for speech and language offerings has become more mature,
evidenced by an increasing emphasis and interest in end-user applications,
enterprise solutions and consulting and services."
L&H is one of several companies offering "continuous speech" PC-based
software and it has attracted funding from both Intel and Microsoft, which
plans to incorporate speech into future products.
Other contenders in this fast-growing market include Dragon Systems, a US
pioneer, and mainstream IT vendors such as Philips and International
Business Machines - the market leader in the US consumer market - which
alone has more than 200 people dedicated to research, product
development and marketing of speech technology.
Earlier this year, Microsoft exercised its option to increase its investment in
L&H to about 7 per cent, while Intel announced plans to invest Dollars 30m.
L&H currently has more than 1,000 people dedicated to the development of
speech technology.
Today's PC speech programs have claimed accuracies of 95 per cent and
recognise continuous speech, so eliminating the need to pause between
words - a big drawback of early systems. "Speech technology is ready to
become a critical element of the PC user interface - and for good reason,"
notes Intel, the semiconductor group.
"Speech adds tremendous value to a wide variety of application and, for
example, can help businesses reduce training and customer support costs,
raise employee productivity, and support international, web-based electronic
commerce."
Several technology developments have helped spur interest in speech
technology. These include a steady increase in desktop computing power,
most recently Intel's Pentium III microprocessor, developments such as the
Universal Serial Bus (USB) and faster dynamic random access technologies,
improvements in speech algorithms and advances in signal processing.
As a result of breakthroughs in these and other technologies, there have
been rapid improvements in the quality and affordability of speech
applications. Just a few years ago, these applications required specialised
hardware and cost thousands of dollars. They had a highly limited
vocabulary, only worked with discrete as opposed to continuous speech,
and required a significant "enrolment" period to train the product to
recognise the user's individual voice characteristics.
Today, some commercial speech applications have broken the Dollars 50
price barrier, offer an extremely high accuracy and a vocabulary of more
than 100,000 words, and require no special-purpose, dedicated hardware.
Dragon Systems' Naturally Speaking package includes a 230,000-word
vocabulary and can handle dictated speech at up to 160 words a minute.
Further improvements in the next few years in core technologies and natural
language processing will increase the cost effectiveness of speech
applications as well as their ability to process natural language.
L&H has introduced several advances this year to make speech technology
much easier to use and more accessible. These include RealSpeak, L&H's
natural sounding text-to-speech technology, and XpressStart, which cuts the
time required to train the group's flagship VoiceXpress package to only five
to 10 minutes. |
