说明：  语音的端点检测是根据语音的特征来进行了，本文根据语音不同帧之间能量的不同，一帧一帧的进行比较
(Speech endpoint detection is based on the characteristics of voice, different voice between the different frames of energy, a one to compare)