摘要 |
PROBLEM TO BE SOLVED: To allow a voice synthesizer, which synthesizers voice form a line of letters, to provide a synthesized voice that sounds like a natural expression. SOLUTION: A word identifying mechanism 10 cuts words out of a line of letters and specifies a line of KANA letters that is to be read. A paragraph judging part 101 performs a judgment process for judging paragraphs included in the line of letters for voice synthesis. A rhythm specifying mechanism 11 specifies a rhythm for the line of letters. In that case, an emphasis effect processing part 111 adds the effect of emphasizing the voice for synthesis to the starting point of a portion that has been judged as a paragraph. A voice synthesizing mechansim 12 synthesizes a voice for output. Since the synthesized voice output is emphasized at the starting point of the paragraph portion in such a way that it is read strongly or slowly or with a breath taken, the periods of sentences become clearer, and the sentences can be read with an expression that is as natural as that by humans. |