| 摘要 |
PURPOSE:To render the change of vocal sounds duration length according to pronouncing speed with fidelity by storing plural constant tables used in the setting of the vocal sounds duration length, and using the constant tables set differently based on requested pronouncing speed selectively. CONSTITUTION:The parameter system of a voice element piece prepared in advance is read out at a rule voice synthesizer according to an input symbol string which expresses a vocal sounds and rhythm, and a voice parameter sys tem is connected with a voice parameter coupling rule. Rhythm is attached by calculating the vocal sounds duration length corresponding to the input symbol string which expresses the vocal sounds and the rhythm with a rhythm control rule, a fundamental frequency pattern, and an amplitude pattern, etc. The plural constant tables used for the setting of the vocal sounds duration length are stored, and the vocal sounds duration length is calculated by using the constant tables set differently based on requested utterance speed selectively, therefore, the change of the vocal sounds duration length due to the change of utterance speed can controlled precisely be. |
