Speech recognition experiments in a noisy environment using auditory system modelling