摘要 |
PROBLEM TO BE SOLVED: To impart proper expression to a synthesized speech vocalizing document data aloud through speech synthesis when the synthesized speech is output.SOLUTION: In a speech synthesis process, specified document data is analyzed so as to derive a text expression distribution for each character (S310-S350). Sound source data corresponding to casting information is acquired and analyzed to derive a sound source expression distribution for each sound source data (S360, S370). Further, a speech parameter is derived by correcting a speech parameter so that the sound source expression distribution matches the text expression distribution for each character (S380, S390). Speech synthesis of a document represented with the specified document data is performed using the corrected parameter and a synthesized speech is output from a speech output terminal (S400, S410). |