The long-term objective of this research is the development of a one-hand controlled speech synthesizer, to provide the user with real-time speech synthesis having a better combination of intelligibility, naturalness, and production speed than current text-to-speech systems. Such a hand synthesizer could give laryngectomees and other speech persons a means of producing higher-quality speech with less effort than other currently available methods (e.g., an electrolarynx, or typing with a text-to-speech system). In Phase I, a hand-held device similar to a pen connected to an articulated arm for measuring six degrees of freedom (three Cartesian and three rotational dimensions) will be the user interface to an HLsyn-based speech synthesizer. With this interface, the user will control the HLsyn parameters for pitch, the first three formants, subglottal pressure and glottal area. Two switches will be added to the interface to control velar and alveolar constriction and occlusion. Subjects will be trained and evaluated on the system progressively, by starting with four of the control parameters and gradually adding the other parameters one at a time. Evaluating the intelligibility, naturalness, and speed of the hand-produced speech will start with single-syllable words containing vowels, diphthongs and semivowels (using pitch and the first three formants), then will include the aspirate/h/and glottal stop (using subglottal pressure and glottal area), and finally will include the consonants/g,k,d,t,z,s/(controlled by the two switches). Additional controls for the remaining phonemes, and device portability will be studied in a subsequent Phase II grant period.
