Spectral Analysis of Noise Corrupted Speech using Auditory Models