论文部分内容阅读
最小均方误差幅度谱的算法,是语音增强算法中的常用算法,但是语音是随机信号,传统的最小均方误差算法不能跟踪实际环境中噪声及频谱增益的变化,从而使性能逐渐下降。本文根据噪声及频谱特性提出了一种改进的最小均方幅度谱算法。该算法结合噪声谱估计并对频谱增益进行了修正,分别计算两种算法信噪比的提高值,并在matlab中适当修改参数后仿真实现。结果表明该算法能有效地抑制背景噪声,提高信噪比。
The algorithm of minimum mean square error amplitude spectrum is a commonly used algorithm in speech enhancement algorithms. However, speech is a random signal. The traditional minimum mean square error algorithm can not track the change of noise and spectrum gain in the actual environment, so that the performance decreases gradually. In this paper, an improved minimum mean square amplitude spectrum algorithm is proposed based on noise and spectral characteristics. The algorithm combined with noise spectrum estimation and spectrum gain correction, respectively, the two algorithms to improve the value of the signal-to-noise ratio, and matlab in the appropriate parameters to modify the simulation. The results show that this algorithm can effectively suppress background noise and improve signal-to-noise ratio.