The area of speech recognition has slowly been growing. However, there
have been many major accomplishments. Microsoft has been working on a program
that will work on many existing PC platforms. Their program,
Whisper, attempts to synthesize many different aspects of present speech
recognition technologies. The different classifications involved in a speech
recognition system are:
- Speaker dependence--whether the system is based on one voice
or can be applied to many different voices.
- Word speed--whether the system can recognize words on a
discrete (i.e. One / word / at / a / time) or continuous (normal conversation).
- Vocabulary size--how many words can the system identify.
Whisper takes all factors into consideration. The program itself combines many
different aspects of computer science, as well. For example in the
"dictionary" that it uses, there are many different types of searches to help
increase its efficiency. In doing this, it can recognize words at a much
quicker pace.
However, rather than have you read through all of this document, here is a
definition of what the links below link to.
- Top--brings you back to this page.
- How it Works--tells you how a system works.
- Useful Links--gives you a list of useful links to learn about current projects.
- Bibliography--gives you a list of the sources that I used.