Auditory prostheses in realistic listening environments

University Of California Irvine, Irvine CA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Course spectral information sufficient for most Cl users to follow conversations in quiet environments, but performance deteriorates substantially with background noise, particularly if the noise is also speech. Assessing the relative importance of speech envelope and spectral fine structure to speech perception in noisy backgrounds is critical for the development of new signal processing strategies for cochlear implant users. Two main experiments are proposed to investigate auditory object identification and spatial segregation for speech (T) presented in noise (M). Experiment 1a examines three types of masking of a target sentence: masking release (using amplitude-modulated noise), energetic masking (using steady-state noise), and informational masking (using competing speech) in Cl and normal hearing (NH) listeners. Envelope cues play a dominant role here, but spectral fine structure becomes more important when segregating competing voices based on differences in voice pitch. Experiment 1b measures pitch differences necessary for segregating competing voices by varying the fundamental frequency (F0) of the competing speech. Experiment 2a compares "poor ear" (T to deaf ear, M to good ear) and "better ear" (T to good ear, M to deaf ear). Envelope cues suffice for these S/N comparisons. However, fine structure will be necessary for localization tasks based on interaural time differences in bilateral implant simulations (Experiment 2b).