Ear Modeling and Sound Signal Processing

Ear modeling can significantly improve sound signal processing and the design of hearing devices. Ear models based on mechanics and neural phenomenology of the inner ear (cochlea) form a class of nonlinear nonlocal dispersive partial differential equations (PDEs). These PDEs capture the essential nonlinear phenomena in hearing, such as combination frequency generation, compression, and suppression, in agreement with experimental data. The PDEs help to generate linear and nonlinear auditory transforms, including a class of orthogonal discrete auditory transforms with broader sprectrum than that of the discrete Fourier transform, consistent with human audition. The PDEs also facilitates an alternative sound amplification method for hearing aid design. ∗Department of Mathematics and ICES, University of Texas at Austin, Austin, TX 78712. Email: [email protected]. Partially supported by NSF-ITR 0219004 and a Guggenheim fellowship.