Articulatory movement formation by kinematic triphone model

The method described predicts the trajectories of articulatory movements by using a kinematic triphone model and the minimum-acceleration model. The kinematic triphone model, which is constructed from articulatory data obtained in experiments using a magnetic sensor system, is characterized by three optimized kinematic features for a triphone and by the intervals between two successive phonemes in the triphone. After a kinematic feature for a phoneme in a given sentence is extracted, the minimum-acceleration trajectory which coincides with the extremum of the time integral of the square of the magnitude of acceleration of the point is formulated. The calculation of the minimum-acceleration requires only linear computation. The method predicts both the qualitative features and the quantitative details experimentally observed.