Auditory modelling accuracy trades off sharply against computational cost

Evidence

This chapter reviewed different computational models of the auditory system, covering a wide range of modelling principles that aim to emulate the processing occurring within different regions of the auditory pathway in varying detail. The models based on windowing the ear canal signals and performing DFT-based processing can be used to explain the basic properties of auditory frequency resolution. Although DFT-based modelling has some drawbacks in temporal accuracy, it is interesting in the scope of this book, as many perceptual audio-coding methods are based on similar processing. The auditory models that model the cochlea using filter banks are more precise, and the best accuracy of peripheral modelling is obtained with transmission-line models. Unfortunately, the computational complexity increases drastically when the accuracy of modelling is increased. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 1741

Categorization

• XREF: Connects to perceptual audio-coding (MP3/AAC psychoacoustic models) which use DFT-based processing — bridges hearing science to consumer audio technology.