摘要 |
PROBLEM TO BE SOLVED: To perform voice synthesis with respect to utterance of a user as if it is a conversation with a person.SOLUTION: A voice synthesizer comprises a voice input section 102 to which utterance by a speaker is input, a pitch analysis section 106 for analyzing a pitch in a specific first section in the utterance, an answer creation section 110 for acquiring an answer with respect to the utterance, a voice synthesis section 112 for performing voice synthesis on the acquired answer by predetermined agent attribute, and a voice control section 109 for controlling voice synthesis of the voice synthesis section 112 with a rule of changing a pitch in a specific second section in the answer to become a pitch having a predetermined relation with respect to the pitch in the first section and correcting the rule according to at least one of speaker attribute of the speaker and the agent attribute. |