What if we just could get rid of our keyboards, and
treated the computer as a normal human, communicating by speech? A
futuristic dream? No it isn't.
Speech recognition can make this dream come true.
Several companies have been doing research on text-to-speech and speech recognition programs, each producing their own programs.
Not everyone is good at typing, and just talking to your computer is a much more natural way of communicating: it's faster, easier... The first speech recognition programs were quite limited: each word had to be trained before it could be recognized, meaning you had to learn the program to recognize each word. Of course, the program was unable to translate speech to text. It was impossible to train each word that appeared even in a short text. The program could only be used to give simple commands like "close window" or "open wordperfect". After a while, new programs appeared which could recognize a quite big vocabulary, without the need to train each word. But it was still not possible to talk naturally to the program. While dictating texts, you had to pause a short while between each word, letting the program a little time to process the word. This method is called discrete speech.
The newer programs are able to recognize continuous
speech, but it is required that you read a few texts to train the
program. The program analyzes your speech and adapts itself to your
way of speaking, so it can produce better results.
The general rule is: the more training, the better results.
Think about the possibilities of such a program: you can dictate texts, without the need to use your hands! It's much faster too: did you ever try to type as fast as you can speak? Trust me: it's very difficult. Speech recognition makes it possible. Another advantage of speech recognition-programs is that they automatically translate words that are typical for our speaking language in written expressions. For example: If you say an time like "half past seven" the computer changes that to "7.30". The same goes for numbers: you say "three hundred and seventeen", the computer notes: "317".
Does it work?
Text-to-speech: Talking Computers!