摘要 |
<p><P>PROBLEM TO BE SOLVED: To provide a high quality voice synthesis apparatus for generating a voice data of a sentence composed of a fixed part and a variable part, by combining recorded voice with rule synthesis voice, in which discontinuity of tone quality and prosody is not perceived when the recorded voice and the synthesis voice are connected. <P>SOLUTION: The voice synthesis apparatus comprises: a recorded voice storing section 5 for storing the recorded voice data including the recorded fixed part in advance; a rule synthesis section 7 for generating a rule synthesis voice data including the variable part and at least a part of the fixed part, from a received text; a conjunction border calculation section 8 for calculating a conjunction border position in an overlapping section of the recorded voice data and the rule synthesis voice data, based on sound feature information of the recorded voice data and the rule synthesis voice data, which corresponds to the text; and a conjunction synthesis section 9 for generating the synthesis voice data which corresponds to the text by connecting the recorded voice data and the rule synthesis voice data which are divided and segmented at the conjunction border position. <P>COPYRIGHT: (C)2008,JPO&INPIT</p> |