| 摘要 |
PURPOSE:To enable a user to stop speaking for an optional time by providing a storage means which stores duration independently set at a place in information, prescribing the word order of the vocalization where the vocalization is possibly stopped halfway. CONSTITUTION:A speech recognizing process part 2 receives a VQ code sequence from a speech signal process part 1 and performs a searching process for recognizing a continuously vocalized speech by referring to a phoneme model storage part 3, a word dictionary storage part 4, and a vocalization grammar storage part 5. An interaction control part 6 interprets a user's request from the recognition result from the speech recognizing process part 2, and generates an instruction to equipment 7 and vocalizes deficient information to the user through a sentence generation part 8 and a speech synthesis part 9, thereby prompting the user to offer information. Further, a continuance length is set to a state wherein the vocalization is possibly stop halfway. The when the vocalization is stopped continuously for longer than the set duration, the vocalization analysis result up to the point of time is outputted. Therefore, the user can stop speaking for an optional time. |
