Vocal emotion identification by older listeners with hearing loss

Abstract

A speaker’s emotional state is one important type of information carried by the speech signal. Numerous studies have been conducted on younger and older normal-hearing listeners’ ability to identify vocal emotions. Much less is known about how hearing loss affects emotion identification, and whether listeners with hearing loss use similar acoustic cues as normal-hearing listeners to identify emotions. The purpose of this study was to evaluate vocal emotion identification performance by younger normal-hearing listeners and older listeners who use hearing aids, and to examine possible acoustic cues used by listeners to distinguish between emotions. The results showed that older listeners with hearing loss performed much worse in a quiet listening environment (38.9%) than young normal-hearing listeners in quiet (84.7%) or in noise (68.4%), and that the use of hearing aids did not improve their performance significantly. Consideration of the pattern of identification errors and the acoustics of the stimuli suggested that the two listener groups relied on different F0-related cues to distinguish emotions: young listeners related on F0 mean and F0 contour, while older listeners relied mainly on F0 mean. Understanding how listeners with hearing loss identify emotion would provide guidance for developing rehabilitative strategies for hearing loss.