论文部分内容阅读
针对非平稳环境噪声提出一种基于噪声整形的语音去噪算法。该算法以最小感知均方误差为准则,在Wiener滤波的基础上,采用听觉感知加权函数修正Wiener滤波方程,实现对噪声谱整形,使噪声谱分布特性跟随语音谱而变;同时引入频率补偿因子克服非平稳噪声谱对语音影响的不均匀性;采用快速噪声估计算法实现对非平稳的估计。实验表明,该算法能更有效地抑制背景噪声,提高了去噪后的语音质量。
Aiming at non-stationary environmental noise, a speech noise reduction algorithm based on noise shaping is proposed. Based on the Wiener filter, the Wiener filtering equation is modified by using the perceptual weighting function to realize the noise spectrum shaping, which makes the noise spectrum distribution follow the speech spectrum. At the same time, the frequency compensation factor Overcoming the inhomogeneous influence of nonstationary noise spectrum on speech; and applying fast noise estimation algorithm to realize non-stationary estimation. Experiments show that this algorithm can effectively suppress the background noise and improve the speech quality after denoising.