A Physiological Ear Model for the Emulation of Masking

Abstract

Auditory masking is exploited in technical audio systems for transmission and storage to improve the perceived audio quality. In this respect, perception models are necessary for a masked threshold prediction to quantify the threshold of audibility for a given signal or distortion. The presented ear model emulates those parts of the auditory sound processing which are relevant to masking. The physiological modeling approach is not limited to simple masker-test-signal configurations as usually employed in psychoacoustical measurements, but it is valid for complex signals as well, which is desired in the aforementioned applications. The ear model includes a nonlinear active cochlear model which emulates the basic mechanisms effective in masking: excitation and suppression. Nonlinear mechanical amplification by means of the outer hair cells plays a crucial role in proper modeling of these effects. In addition to physiological measurements at the mechanical level of sound processing in the cochlea, classical masking patterns and nonlinear masking effects are demonstrated by the ear model.