运用MMSE先验信噪比估计进行语音增强

摘要

针对DD (Decision-Directed)先验信噪比估计方法在处理语音时产生延迟以及非因果先验信噪比估计算法不具实时性的缺点,提出一种MMSE(Minimum Mean Square Error)先验信噪比估计方法.它在高斯语音模型假设的基础上,运用最小均方误差准则直接从带噪信号中估计先验信噪比.通过对增强语音信噪比、Itakura-Saito失真测度以及信号时域图和语谱图仿真,结果表明,该算法比DD算法能更好地抑制“音乐噪声”和防止语音畸变,且相对于非因果先验信噪比估计算法具有更强实时性.%To solve the problems that Decision-Directed(DD) a priori SNR estimation gets defects of delayed and non-casual a priori SNR estimation has shortcoming of non-real time.MMSE a priori SNR estimation is proposed in this paper.Based on the Gauss voice model assumption, it directly estimates SNR of noisy speech using the Minimum Mean Square Error (MMSE) methods.Simulation with the SNR.Itakura-Saito distortion measure,signal wave and spectrogram show that the proposed method is not only better than DD in the area of suppressing "musical noise" and preventing the voice distortion but also superior to non-causal a priori SNR estimation on real-time.