| 摘要 |
PURPOSE:To enable speech synthesis corresponding to a dialect and to obtain a synthetic speech of high quality with high understanding by providing a dictionary which contains pieces of information corresponding to dialects and specifying information used for synthesis. CONSTITUTION:A dialect type setting part 11 sets the information used for the synthesis and an input document is divided by a text analysis part 2 to obtain information on an accent type and reading. Then a rhythm processing part 5 determines a pause position, the accent type of a paragraph, and the intonation of the whole sentence and a phoneme processing part 6 obtains actual speech inscription; and a synthesis parameter generation part 7 obtains actual synthesis parameters and a synthesis part 12 synthesizes a speech by using a formant type synthesizer. Thus, plural pieces of information corresponding to the dialects are contained in the dictionary and it is decided which dialect type information is used for the synthesis to improve the understanding of the synthetic speech and also obtain the synthetic speech which is familiar and natural. |
