Non-Stationary Noise Power Spectral Density Estimation Based on Regional Statistics

Xiaofei Li (1) , Laurent Girin (2, 1) , Sharon Gannot (3) , Radu Horaud (1)

1 PERCEPTION - Interpretation and Modelling of Images and Videos
2 GIPSA-CRISSP - GIPSA - Cognitive Robotics, Interactive Systems, & Speech Processing
3 Faculty of Engineering [Israel]

Estimating the noise power spectral density (PSD) is essential for single channel speech enhancement algorithms. In this paper, we propose a noise PSD estimation approach based on regional statistics. The proposed regional statistics consist of four features representing the statistics of the past and present periodograms in a short-time period. We show that these features are efficient in characterizing the statistical difference between noise PSD and noisy speech PSD. We therefore propose to use these features for estimating the speech presence probability (SPP). The noise PSD is recursively estimated by averaging past spectral power values with a time-varying smoothing parameter controlled by the SPP. The proposed method exhibits good tracking capability for non-stationary noise, even for abruptly increasing noise level.

Mots clés: regional statistics; speech presence probability; noise PSD

Domaines: Acoustique [physics.class-ph]; Traitement du signal et de l'image [eess.SP]; Son [cs.SD]

Xiaofei Li, Laurent Girin, Sharon Gannot, Radu Horaud. Non-Stationary Noise Power Spectral Density Estimation Based on Regional Statistics. ICASSP 2016 - 41st IEEE International Conference on Acoustics, Speech and Signal Processing, IEEE Signal Processing Society, Mar 2016, Shanghai, China. pp.181-185, ⟨10.1109/ICASSP.2016.7471661⟩. ⟨hal-01250892⟩