| 基于听觉掩蔽效应和噪声估计的语音增强算法 |
| |
| 引用本文： | 阚仁根,于凤芹.基于听觉掩蔽效应和噪声估计的语音增强算法[J].电声技术,2008,32(2):55-57,66. |
| |
| 作者姓名： | 阚仁根 于凤芹 |
| |
| 作者单位： | 江南大学,通信与控制工程学院,江苏,无锡,214122 |
| |
| 摘 要： | 谱减法在增强语音、提高信噪比的同时,残留的音乐噪声较大.在利用听觉掩蔽闻值对谱减系数进行修正的基础上,采用实时噪声估计来减少谱减法噪声估计误差,并对谱减后的语音信号进行感知滤波来进一步抑制残留音乐噪声.实验结果表明,该算法能去除噪声,增强语音,并在不影响信噪比的同时降低语音失真测度值.主观测听表明语音音质有明显提高.
|
| 关 键 词： | 语音增强 听觉掩蔽 噪声估计 感知滤波器 听觉掩蔽 掩蔽效应 噪声估计 语音增强算法 Noise Estimation Properties Auditory Masking Human Based Enhancement 语音音质 主观测听 测度值 语音失真 高信噪比 影响 去除噪声 结果 实验 滤波 |
| 文章编号： | 1002-8684（2008）02-0055-03 |
| 收稿时间： | 2007-11-17 |
| 修稿时间： | 2007年11月17 |
Speech Enhancement Based on Human Auditory Masking Properties and Real-time Noise Estimation |
| |
| KAN Ren-gen,YU Feng-qin.Speech Enhancement Based on Human Auditory Masking Properties and Real-time Noise Estimation[J].Audio Engineering,2008,32(2):55-57,66. |
| |
| Authors: | KAN Ren-gen YU Feng-qin |
| |
| Abstract: | The speech enhancement is to enhance the SNR. But when the enhanced speech is obtained, some music noise is left. When the SNR is enhanced, the voice has a distortion. A real time noise estimation in order to reduce the error of the noise estimation is given. During the estimation of power spectrum of the speech, the parameters of estimator can be modified. Finally, the filter of apperception is used to reduce the music noise to the noise masking. The noise is reduced in a high leave and distortion of voice is eliminated in this algorithm. |
| |
| Keywords: | speech enhancement human auditory masking noise estimation filter of apperception |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
