Noise power estimation based on a sequential Gaussian Mixture Model

Noise estimator is a basic component of noise reduction algorithms. In this paper, we propose a sequential Gaussian Mixture Model (SGMM) to track noise power in log-spectral domain. This model comprises two Gaussian components, which respectively describe the speech and nonspeech log-power distributions. The initial distributions are firstly established by EM algorithm, and then sequentially updated in an on-line manner. The mean of the nonspeech component is the noise estimate based on the maximum likelihood. For the sake of reliability, some constraints are introduced to this SGMM. The proposed algorithm is compared with MS, MCRA and IMCRA algorithms, and it shows promising performance.