SampEn是关于时间序列的规律估计近似值的一个变量。其基本思想是基于一个音频的规律性将其插入到语音或者音乐中。通过加窗计算的SampEn序列，所插入的音频流的规律性就被估计出来了。上诉方法的有效性已经被试验证实，这其中包括BBC的新闻广播和不同类型的具有多种暂时分类语音和音乐。结果显示，对于所有测试的实验，上述方法的健壮性达到了准确区分率高于90 。-SampEn time series on the approximation of the laws of a variable is estimated. The basic idea is based on the regularity of an audio plug it in to the voice or music. Calculated by the addition of window SampEn sequence, the insertion of the regularity of the audio stream has been estimated that out. Effectiveness of the method of appeal has been confirmed that the pilot, including the BBC' s news and a variety of different types of temporary classification of voice and music. The results showed that the test for all experiments, the robustness of these methods to achieve an accurate distinction between the rate of higher than 90 .