Catalogue of Artificial Intelligence Techniques
Author(s): Steve Isard
A speech signal can be described through formants; these are the principal resonances of the vocal tract, that determine which overtones of the fundamental larynx frequency will be allowed to pass through into the speech signal, and which will be damped. Each position of the articulators (the lips, tongue, jaw, etc.) gives rise to a characteristic pattern of formants. A speech signal can therefore be represented in terms of formant patterns which change over time, both for purposes of recognition and synthesis-by-rule.
- Markel, J.D. and Gray Jr., A.H. , Linear Prediction of Speech
, Springer-Verlag, New York, 1976.