C H A P T E R 2
Using Dragon NaturallySpeaking Successfully
Dragon NaturallySpeaking User’s Guide
8
Dragon NaturallySpeaking uses acoustical data about the sound patterns
that make up different words to choose the words that most closely
match what it heard. Since no two people sound exactly alike, Dragon
NaturallySpeaking does a much better job of matching sounds to words
when it knows something about your pronunciation. When you first
trained the program, you provided acoustical data on top of what Dragon
NaturallySpeaking already knows about the sounds of English.
Sometimes it’s not possible for Dragon NaturallySpeaking to choose the
correct word based on sound alone. Consider these two phrases: “pizza
delivery boy” and “Pete’s a delivery boy.” When spoken, they sound
exactly alike. How would Dragon NaturallySpeaking know which to
choose? When it’s not clear from sound alone what was said, Dragon
NaturallySpeaking uses language data about the frequency with which
words in English appear alone and in combination with others to
determine which words were most likely spoken. The phrase “pizza
delivery boy” is more common than “Pete’s a delivery boy,” so the
program would favor this phrase over the other.
Since people write differently, it helps if Dragon NaturallySpeaking
knows something about the frequency with which you use different
words. When you run Vocabulary Builder™, or make corrections in the
Correction dialog box, you’re providing the program with language data
about how often you use different words.
Knowing that Dragon NaturallySpeaking uses acoustical data and
language data to recognize your speech can help you know what to do to
make the program work better.
ugpro.bk Page 8 Monday, August 9, 1999 9:42 AM
