论文部分内容阅读
语音增强效果的提高,有赖于对噪声的准确估计和对噪声变化的及时跟踪与更新。为了提高对非平稳噪声的估计和更新能力,本文基于“改进的最小值控制递归平均”(IMCRA)算法,提出了噪声谱最小值双向搜索的改进算法。该算法结合前向搜索和后向搜索谱最小值方法的特点,有效提高噪声估计的准确性、减小非平稳噪声跟踪的延迟。实验仿真表明:在非平稳噪声环境和低信噪比条件的语音信号增强处理中,本文提出的改进算法非常有效,与IMCRA算法相比,它可以获得更好的分段信噪比的提高。
The enhancement of speech enhancement depends on accurate estimation of noise and timely tracking and updating of noise changes. In order to improve the estimation and updating ability of non-stationary noise, this paper proposes an improved algorithm of bi-directional search based on the minimum recurrent mean of noise (IMCRA) algorithm. Combining the characteristics of forward search and backward search spectrum minimum, this algorithm can effectively improve the accuracy of noise estimation and reduce the delay of non-stationary noise tracking. Experimental results show that the improved algorithm proposed in this paper is very effective in non-stationary noise environment and low signal-to-noise ratio speech signal enhancement. Compared with the IMCRA algorithm, it can achieve better segmentation signal-to-noise ratio improvement.