We have covered a few specific application examples in detail in previous chapters, to illustrate some of the ways that an auditory-image front end can be connected to a higher-level application system. In this chapter, we very briefly survey some of the other areas where front ends based on models of hearing have been used to advantage, and where further advances are to be expected.

Auditory Physiology and Psychoacoustics

As the chapter-opening quote by Karjalainen suggests, problems in the physiology and psychophysics of hearing, such as those introduced in Chapter 4, can be addressed most fruitfully in the context of computational models. Simple models, such as the Fourier spectrum view of sound, have been useful historically, but they run into limits and lead to questions that can only be addressed in the context of more realistic models. Progress in understanding details of psychoacoustic effects such as loudness, masking, pitch, timbre, etc. have come about gradually as these phenomena have been studied in the context of increasingly detailed models of auditory physiology, especially including cochlear function.

An example is the study of auditory filter models for explaining simultaneous masking, as discussed in Chapter 13. Early auditory filter models were simple functional bandpass concepts (rectangular, Gaussian, and simple resonance filters). Later models incorporated more accurate and flexible filter shape descriptions and level-dependent parameter changes. Most recently, we showed that filter models derived from active wave propagation concepts lead to better fits to the data with fewer adjustable parameters (Lyon, 2011b,a). In this sense, using more knowledge of the physiology has led to better explanation of the psychophysics, and thereby a reinforcement of models of both.