摘要 |
<p>PURPOSE:To attain a stable recognizing rate while holding an optimum sound volume for recognition by supervising the sound volume inputted to a voice input device and feeding back the result to an operator by display or voice. CONSTITUTION:A voice signal collected by an acoustoelectric transducer 1 is inputted to a recognizing section 5 via an amplifier 2, an AGC circuit 3 and a voice cutout section 4. The output of the amplifier 2 is inputted to a sound volume detecting section 6 at the same time to detect abnormal amplitude. Further, the output of the AGC circuit 3 is inputted also to the sound volume detection section 6. The sound volume detecting section 6 measures whether or not the maximum amplitude and the minimum value of the energy are within a prescribed level in the voice section, transmits a signal to a display section 9 and displays it on a display 10. If they are not within a prescribed level, the operation of the recognizing section is interrupted and also it is informed by the display 10 and a receiver 8.</p> |