Voice input - General description

There are two ways of voice recognition:

speaker independent

- computer learns, how people pronounce things in a particular language, and then recognizes a phrase by comparing it to its "dictionary" of phrases;

speaker dependent

- computer records few samples of a phrase spoken by a particular speaker, and then recognizes it when it is spoken again.

The great advantage of speaker independent way is, that you don't have to do any setup - you just sit by a computer and talk to it. There are few very important disadvantages though:

Therefore I chose for KLearnNotes2 the second option: speaker dependent model.

The price to pay is that starting with this would require that each user makes some setup. I expect it to take (including reading the quick start instructions) less then 5 minutes for a novice, unexperienced user to make first voice-answered test. :)