摘要 |
PROBLEM TO BE SOLVED: To improve the operability at the time of speech input manipulation by judging whether the speech to be recognized is inputted in an adequate state or not by the magnitude of the speech power within an effective speech section and emitting the information indicating adequateness right after the input end of the speech to be recognized when judgment is made that the input is adequate. SOLUTION: A speech section deciding section 13 detects an effective speech section by using the speech power calculated in a speech analyzing section 12. If the speech section exists inside the range of the two thresholds having the max. value of the speech power, it is judged that the effective speech section is within a normal range and the speech is judged to be inputted in the state adequate for execution of the speech recognition. When the judgment is made that the speech input is adequate, the signal indicating that the speech is inputted in the adequate state is output after a specified time from the terminal of the effective speech section. An information output section 15 forms the response speech corresponding to the result of the recognition and a speech message of the content to ask a question to a user by using data 17 for output and outputs both. |