ABSTRACT
The use of auditory models as speech recognition front ends has recently attracted a great deal of interest. The underlying assumption is that a good model of the auditory system should generate a more natural and efficient representation of speech compared to ordinary spectrum analysis. However, we have to keep in mind that only some of the peripheral processes of sound perception are included in most existing models. It is claimed that three aspects of the acoustic-to-auditory transformation most necessary "to include in any new front-end designs of speech recognizers." The three effects are the mechanical filtering in the cochlea, the mecano-electric transduction, and the temporal adaptation.
