| 摘要 |
PROBLEM TO BE SOLVED: To perform synthesis complying with a dynamic request for a target generation time for a synthesized speech, a load on a central processing arithmetic unit of a speech synthesizer, or quality of the synthesized speech etc., at the request. SOLUTION: The speech synthesizer comprises a memory stored with compressed elementary speech units, uncompressed elementary speech units corresponding to the compressed elementary speech units, or difference elementary speech units based upon the compressed elementary speech units and uncompressed elementary speech units corresponding to the compressed elementary speech units and the compressed elementary speech units, an elementary speech unit selection part which selects elementary speech units stored in the memory, and an elementary speech unit generation part which reads one of compressed elementary speech units and uncompressed elementary speech units out based upon the choice made by the elementary speech unit selection part. COPYRIGHT: (C)2006,JPO&NCIPI
|
