Fig. 3From: Bayesian STSA estimation using masking properties and generalized Gamma prior for speech enhancementVariation of the proposed choice of α versus frequency compared to that of initial speech STSA estimate for one sample time frame of the noisy speechBack to article page