| 摘要 |
PURPOSE:To curtail the processing quantity for segmenting a voice section by measuring a continuation time of a high level signal being higher than a threshold level and a continuation time of a low level signal being lower than the threshold level and deciding the voice section, based on the continuation time. CONSTITUTION:A single threshold level and an input sound signal are compared by a comparing means 1, and when said signal level is higher than the threshold level, and when said signal level is lower, a high level signal and a low level signal are outputted, respectively, a continuation time of both the signals is measured by a measuring means 2, and when the high level signal is continued exceeding the first prescribed time, a deciding means 3 decides its rise time point to be an initial end. On the other hand, when the low level signal is continued exceeding the second prescribed time, its fall time point is decided to be a terminal end. In the case when the continuation time of the first high level signal is shorter than the first prescribed time, and the continuation time of the low level signal is shorter than the second prescribed time, and the next high level signal is continued exceeding the first prescribed time, the rise time point of the first high level signal is decided to be an initial end. In such a way, the processing quantity for segmenting a voice section can be curtailed. |
