| 摘要 |
PURPOSE:To obtain the voice response device being free from a sense of incongruity by providing a sound recording and editing part, a parameter retrieving part, a character-train analyzing pat, a control part and a parameter interpolation coupling part. CONSTITUTION:A control part 5 reads out and analyzes voice response message information, sends information related to a fixed message part to a sound recording/editing/ synthesizing part 2 or a parameter retrieving part 3, and also, sends information related to a variable message part to a character-string analyzing part 4. In the synthesizing part 2, a composite voice is generated, based on the information. In the retrieving part 3, a parameter for analyzing and synthesizing is retrieved, based on the information, and in the character-string analyzing part 4, a character-string for expressing the variable message part is analyzed and phoneme information and meter information are generated. A parameter interpolation coupling part 6 brings a voice element piece parameter to interpolation coupling, based on the phoneme information, and generates a phoneme parameter. Subsequently, information related to a sound source of a voice is inputted to a composite filter 10, a digital signal determined by various parameters is brought to D/A conversion 8 and a composite voice is obtained. According to this constitution, a listener has no sense of incongruity. |
