Microscopic prediction of speech recognition for listeners with normal hearing in noise using an auditory model.