Before an engine can begin recognizing speech, the application must inform the engine what it expects the user to speak. To do this, the application uses one of the following kinds of grammars:
· A context-free grammar uses rules that predict the next words that might possibly follow the word just spoken, reducing the number of candidates to evaluate in order to recognize the next word.
· A dictation grammar defines a context for the speaker by identifying the subject of the dictation, the expected style of language, and what dictation has already been done. Not all engines support dictation grammars.
· A limited-domain grammar does not provide strict syntax structures, but does provide a set of words to recognize. A limited-domain grammar is a hybrid between a context-free grammar and a full dictation grammar. Not all engines support limited-domain grammars.
For purposes of the following discussion, the application is using a context-free grammar for command and control.