A system-theoretical approach to modeling the auditory system typically is confined to modeling the essential functional features, and it abstracts to any possible extent from physical and physiological details. What regards the auditory periphery, within the system-theoretical approach it is not necessary to include a detailed model of the cochlea.
However, there obviously is no way to model the auditory system without having a "black box" at the periphery of each ear that has one input gate for the sound stimulus, and a multitude of output gates as its output for subsequent hierarchical processing. In between there must exist a frequency selective system that assigns to each of the output channels a "best frequency", and between the input gate and each output gate there must exist a cochlear transmission function CTF that resembles that of a bandpass filter. That kind of modeling becomes inevitable with respect to the notion that several basic audiory sensations such as, e.g., spectral pitch and roughness, are intimately dependent on the peripheral mechanism that transforms the sound stimulus into auditory excitation. As a first step to designing that black box, the PET system was developed.
As we can discern from both neurophysiological and psychoacoustical tuning curves, the characteristics of the CTF has a very steep slope toward frequencies above the best frequency, and a flat "tail" toward low frequencies. This characteristic was modeled by the CTF described in ,  p. 258-260, .
Author: Ernst Terhardt firstname.lastname@example.org - Feb. 20 2000