Psychoacoustic filtering for noisy speech enhancement

A new denoising approach is introduced in this paper. It is based on the fact that denoising may be performed by mimicking the human ear function in order to improve the psychoacoustics appearance of speech signal. In the proposed method, the speech signal is decomposed by using a gammatone filterbank in accordance with ERB scale. Spectral attenuation filtering is then applied in each sub-band which is based on continuous noise estimation. In the output of the spectral attenuation filter, masking threshold is calculated by using the Johnston model and then inserted in the Psychoacoustics gain filter. Evaluation tests are realized using objective and subjective criterion such as Perceptual Evaluation of Speech Quality (PESQ) for the objectives scores and mean the quality rating of signal distortion (SIG), noise distortion (BAK) and overal l quality (OVRL) for the subjective scores. The results show that our method gives best global quality of the enhanced speech signal while maximizing noise elimination and minimizing distortion.