Skip to main content
Text Size: sML

Second Biennial Hearing Aid Research and Development Conference

Hearing Aid Conference Logo

September 22-24, 1997
National Institutes of Health
Bethesda, Maryland

Cochlear Signal Processing for Compression and Gain Control Extends Dynamic Range and Preserves Temporal Modulation

Julius L. Goldstein, Department of Electrical Engineering, Washington University, St. Louis, Missouri

Cochlear signal processing models developed to represent a broad range of nonlinear cochlear responses (Goldstein, 1990, 1993, 1995; Lin and Goldstein, 1995) suggest that the cochlea provides effective, if not optimal, solutions to the conflicting requirements for signal amplification discovered in hearing-aid research. The need for signal compression to compensate for loss of dynamic range in sensorineural hearing loss (Villchur, 1973) conflicts with the superiority of linear amplification for speech signals within a limited dynamic range (Lippmann et al., 1981). This conflict is explained by the loss of temporal modulation in rapid compression (Plomp, 1988; Drullman et al., 1994). Thus, hearing-aid design guidelines have been based on compromises. Rapid nonlinear compression offers large-range compression (Lippmann et al., 1981), while slower-acting automatic gain control preserves temporal modulation with linear short-term response (Plomp, 1988). Basic cochlear research has shown that normal cochlear operation, with healthy outer hair cells, provides responses that reflect a nonlinear interaction between linear and compressive response mechanisms, in which the compressive gain is potentially controlled by efferent innervation to the outer hair cells (e.g., Kiang et al., 1986). The signal processing models represent normal cochlear response as nonlinearly switching between compressive range-extending responses at low and moderate sound levels, to linear responses at high levels, while efferent gain control has the potential for optimizing the level for switching between the regimes. Although verification that the normal auditory system implements this strategy requires new basic research, the existing models suggest new guidelines for hearing-aid design. Examples will be presented to illustrate the temporal and spectral characteristics of responses from simulated cochlear-model-based designs of nonlinear hearing-aid amplifiers.

[Basic research on cochlear models supported by NINDS and NIDCD]

Return to Conference Abstracts List