Abstract: This paper introduces a novel technique for estimating the signal power spectral density to be used in the transfer function of a microphone array post-filter.The technique is a modification of overestimation of noise power spectral density of the existing Zelinski post-filter and McCowan post-filter，which utilize the auto- and cross- spectral densities of the signals.Near-field compensation is taken into account.In simulations，which make use of Habets’ method of generating isotropic noise for sensor array，is used to test the proposed method.Furthermore，real noisy recordings provided freely by Carnegie Mellon University is also used.Simulation results show that the modified post-filter results in prominent improvement in terms of objective speech quality measures.

Key words: microphone array, post-filter, speech enhancement

摘要： 介绍了一种用于近场麦克风阵列后滤波语音增强的方法，避免了现有的利用自功率谱密度和互功率谱密度的Zelinski和McCowan后滤波器中噪声功率谱过估计的问题，并加入了近场传播的幅度衰减补偿和相位延迟补偿。使用卡内基梅隆大学提供的多麦克风语音数据以及使用Habets E A P提出的生成阵列散射噪声的方法进行的仿真实验，证明了改进的方法在语音质量客观评估量方面优于Zelinski后滤波法和McCowan后滤波法。

关键词: 麦克风阵列, 后滤波, 语音增强