different sites one can visit to do more DD
Saturday, October 17, 1998
Site Seeing
Web wisdom on recognizing spoken words
More FP Technology stories
By GEOF WHEELWRIGHT
For The Financial Post
Star Trek has a lot to answer for. The 1960s hit television show popularized the idea of being able
to talk to computers -- and have them answer back.
Yet it has only been in the past few years that speech technology has been powerful enough to be
used on the average desktop personal computer in any meaningful way. And there is still lots more
work to be done before it achieves the level of widespread use enjoyed by computer keyboards.
This has made it one of the most widely researched technologies in the computer industry and the
academic world. To get a good sense of just where that research is leading, you may want to visit
one of the many World Wide Web sites devoted to this subject:
swww.tiac.net/users/rwilcox/speech.html -- This site, run by speech recognition enthusiast Russ
Wilcox, is not the most technologically taxing side your browser will ever visit, but it is a great
directory for anyone who wants to find out more about the business of speech recognition. There are
links to many software companies, research organizations, university sites, trade show calendars and
online demonstrations. If you want to get a sense of the state of speech technology, this is an
excellent place to start.
shttp://sls-www.lcs.mit.edu/ec-nsf/mit-sls.html -- Your next stop might be at Boston's
Massachusetts Institute of Technology. The spoken language systems group (SLS) at MIT's
computer science lab is devoted to the research and development of what it calls "interactive
conversational systems". Researchers suggest a "speech interface", in a user's own language, is ideal
because "it is the most natural, flexible, efficient, and economical form of human communication."
The MIT group's research activities fall into three categories. In the first, basic research is aimed at
measuring and modeling various aspects of the speech communication chain, ranging from the
development of computational models of how humans hear to the modeling of linguistic regularities in
spontaneous speech.
The group's second area of research will use these results to develop algorithms for speech
recognition and language understanding.
The third research category will integrate the component technologies into prototype spoken
language systems with varying capabilities. Several applications are already under development
including something the group calls its "Voyager" system, designed to help future drivers explore and
navigate in an unknown urban setting, and the "Pegasus" system allowing travelers to make airline
reservations.
swww.bell-labs.com/project/tts/ -- Given Alexander Graham Bell's interest in early speech
technology, it should perhaps come as no surprise that Bell Laboratories (part of Lucent
Technologies) has developed what it calls the Bell Labs Text-to-Speech system (TTS), which is
featured at this site.
Bell Labs suggests its TTS system has various applications, including reading electronic mail
messages, generating spoken prompts in voice response systems, and as an interface to an
order-verification system for salespeople in the field.
In its current form, TTS is implemented entirely in software and only standard audio capability is
required. It also contains several components -- each of which handles a different task. For example,
the text analysis capabilities of the system detect the ends of sentences, perform some rudimentary
syntactic analysis, expand digit sequences into words, and translate and expand abbreviations into
normally spelled words that can then be analysed by the dictionary-based pronunciation module.
Visit this site to hear more details about the project.
swww.research.microsoft.com/research/srg/ -- Microsoft spends more than US$2 billion of
research and development money every year in a number of key areas. Speech technology is one of
them -- and one to which the software giant is heavily committed. At this site, you'll learn about the
work of Microsoft's speech technology group.
Current projects include one code-named Whisper (Speech Recognition) and another that goes by
the name of Whistler (Speech Synthesis).
The group is also working on text-to-speech (where the computer speaks words that you type into
it), spoken language understanding and the development of a "multimodal conversational user
interface."
CANOE home | We welcome your feedback.
Copyright © 1999, Canoe Limited Partnership.
All rights reserved. Please click here for full copyright terms and restrictions.
|
