说明：  一种基于最小均方误差的噪声估计算法，实现语音增强，算法可以直接使用，已调试通过。
(Based on the minimum mean square error of noise estimation algorithm, speech enhancement algorithm can be used directly, have been debug through.)