An evaluation of the properties of a model of the auditory system for speech signal recognition (comparison with linear prediction coding)