Speech Enhancement using a Modified Apriori SNR and Adaptive Spectral Gain Control

A new approach to single channel speech enhancement is proposed using a modified a priori SNR and spectral gain control. The proposed approach is first directed toward finding self adaptive averaging factor to estimate the apriori SNR. Next, spectral gain is reduced in order to suppress effects of the noise in the speech absent frames. Further, in the speech present frames, in order to reduce signal distortion, the spectral gain is controlled to be unity based on an SNR calculated by using a ridgeline spectrum. Finally, the original noisy speech is added to the estimated speech in a ratio is controlled by the long term averaged SNR of the estimated noise and the noisy speech. Computer simulations by using speech signals, the white noise, the car noise and the babble noise have been carried out using several available methods and the proposed method. It is observed that there is improvement in speech quality by the proposed method.