Speech Technology (11-Dec-2003)
Two of the main ways that technology is used are for speech recognition and
speech synthesis.
Speech Recognition
This is typically seen as being useful for dictation, but also has
applications for people who are disabled, or are unable to use their hands for
some reason - e.g. systems have been developed for use in operating theatres
where the surgeon can issue commands verbally while in mid cut. Many of these
systems require training - often needing to learn a particular speaker's
style, and may only understand a limited number of phrases. For those systems
(such as used in dictation) which attempt to understand the whole language,
there are the added difficulties of coping with homophones and punctuation.
One of the main problems is that the sounds of speech are so varied that it is
very difficult to write a program to understand all variations, and it is here
that linguists may be able to offer help.
Speech Synthesis
It is a lot easier to get a computer to talk than to understand, and so there
are many more systems in commercial use which use some kind of speech
synthesis. These may be in situations where it is not practical for someone
to read the information (e.g. directions being given when driving a car), or
where the user is blind (e.g. speaking book). People who have a disability
that prevents them from speaking (e.g. Stephen Hawking) may also make use of
speech synthesis systems. Automated call-handling systems may also use speech
synthesis as a means of saving labour costs. Toys are also a major user of
this type of technology.
Sounds, Grammar and Meaning page