| 摘要 |
PROBLEM TO BE SOLVED: To solve the problem that equipment with a voice interactive function such as a car navigation device and a TV game, apears, however, such equipment does not exist as varies voice quality and tone of output voices according to progress situation of voice interactions, but a synthesized voice is outputted in the same voice quality and tone, therefore, a user loses interest in the machine operation. SOLUTION: Since feelings of a speaker are categorized by analyzing speed or intonation or the like of input voices and responding voices are also categorized correspondingly to the categories so as to respond with feelings, the monotonousness is eliminated, and it becomes possible to construct a user-friendly voice interactive system like a conversation with a human being. Further, responding voices are outputted with more emphasis on responding words corresponding to higher frequent input keywords than those corresponding to lower frequent input words, therefore, the speaker can exactly hear important points among the responding voices. |
