论文部分内容阅读
本文主要研究语音增强系统在短时频域内,以对数最小均方误差作为语音幅值估计的准则.对含噪语音进行消噪处理。首先推导了公式,并从理论上作了分析研究.在此基础上作了大量实验,与类似的最小均方误差短时谱幅值估计法(MMSE STSA)作了性能比较.实验结果表明,由MMSE LOG-STSA增强方法得到的语音输出的质量优于由MMSA STSA方法得到的质量.语音的失真度和清晰度均有显著改善,同时大大降低了语音中的“音乐噪声”.
In this paper, we mainly study the speech enhancement system in the short-term frequency domain, the logarithm of the mean square error as the criterion of speech amplitude estimation. First of all, the formula is deduced and theoretically analyzed. Based on this, a large number of experiments have been done and compared with the similar MMSE STSA method. The experimental results show that, The quality of the speech output obtained by the MMSE LOG-STSA enhancement method is superior to that obtained by the MMSA STSA method, and the speech distortion and sharpness are significantly improved, while the “music noise” in the speech is greatly reduced.