结合短时能量和过零率对一段连续发音的wav文件进行端点检测，将单个词组的发音分割出来。-Combination of short-term energy and zero crossing rate of a continuous sound of the wav file detection, the pronunciation of a single phrase differently.