An Improved LSA-MMSE Speech Enhancement Approach Based on Auditory Perception

Gain function of traditional enhancement algorithm is to estimate every signal spectral component, therefore, this introduce relatively more speech distortion. To improve the effect of speech enhancement at low signal-to-noise ratio (SNR), this paper proposed a optimal speech enhancement scheme. Based on auditory perception properties, no estimator for noise masked spectrum and classical enhancement estimator for noise unmasked spectrum. Then a speech signal estimator is proposed as a weighted sum of the individual estimator in each state, where the weight is related with noise masked probability. Compared with Virag's method and LSA-MMSE estimator, the proposed estimator can suppress the residual noise effectively while keep smaller speech distortion especially at low SNR.