| 摘要 |
<p>PURPOSE:To output a voice whose natural property and intelligibility are good by inserting a voice pause section into an output voice by information from a postpositional word ''ha'' detecting part, etc. and detecting information of a punctuation mark detecting part. CONSTITUTION:A symbol detecting part 17, a postpositional word ''ha'' detecting part 18, a numeral detecting part 22, an in-composite work word boundary detecting part 26, a postpositional word detecting part 23, a character kind changing point detecting part 25, etc. are provided on the outside of a punctuation mark detecting part 19 for detecting a punctuation mark from an inputted Japanese language sentence. Also, when the punctuation mark detecting part 19 has detected a punctuation mark, a voice pause section is inserted into an output voice, and also a pause section setting part 21 inserts the voice pause section into the voice by other detecting information than hat of the punctuation mark detecting part 19. In this way, a natural voice pause section is set automatically, therefore, a voice whose natural property and intelligibility are good can be outputted.</p> |
