To synthesize an aurally natural voice even when a voice element is extended.
A storage device 14 stores a plurality of pieces of voice element data DA each indicating voice element V including a plurality of sound element segments S corresponding to respective sound elements being different from each other. An element control unit 26 performs extention processing for extending an object segment WA so that an extension rate of the central part of the object segment WA is high as compared to those of the head section and tail section, the object segment WA produced by concatenating a sound element segment S2 at the tail of a voice element V1 corresponding to a consonant sound element and a sound element segment S1 at the head of a voice element V2 corresponding to the sound element, and thereby generates synthesized element data DB in an adjustment segment WB having a target time length LB. A voice synthesis unit 28 generates a voice signal VOUT from the synthesized element data DB generated by the element control unit 26.
| JP2004004440A | 2004-01-08 | |||
| JPS62245298A | 1987-10-26 |
Yashiro Hitoshi
Taro Takahashi
Next Patent: VOICE SYNTHESIZER