摘要
提出一种改进的基于听觉掩蔽的自适应阶MMSE语音增强算法,实验表明这种改进算法比起经典的语音增强算法能更显著地提高算法的客观性能,特别在非平稳噪声和低信噪比的环境中能快速估计出变化的噪声功率谱。
An improved masking-based adaptive-order MMSE speech enhancement statistical information noise-estimation used in the Speech enhancement algorithm. Experiments results prove that this algorithm can dramatically improve objective performance compared with the classic algorithm, especially it can estimate the noise spectrum quickly in the non-stationary noise and the low SNR environment.
出处
《电脑开发与应用》
2008年第11期63-64,69,共3页
Computer Development & Applications
