Effect of audibility, frequency region, and hearing loss on better-ear glimpsing

Abstract

Better-ear glimpsing (BEG), which utilises short-term interaural level difference cues (ILDs), helps understanding speech in noise. Since ILDs are mainly available at high frequencies where hearing loss is usually strongest, hearing-impaired (HI) listeners cannot take full advantage of BEG. One possible solution is to provide artificially generated ILDs at low and mid frequencies where hearing loss is usually less pronounced. Rana and Buchholz [JASA, 2016] showed that HI listeners can take advantage of such artificial ILDs, but their received benefit in speech intelligibility was still smaller than for normal-hearing (NH) listeners. To understand how far this difference can be explained by differences in audibility, speech recognition thresholds (SRTs) were measured in four frequency regions (low, mid, high, and broadband) at four audibility levels (0, 10, 20, and 30 dBSL) in 10 NH and 10 HI listeners. SRTs were measured for BKB like sentences in the presence of vocoded speech distractors. The difference between co-located and spatially separated SRTs was used as a measure of BEG. Results revealed that providing equal audibility removed the performance differences between groups. Moreover, the effect of BEG was roughly frequency independent and increased by about 2 dB for each 10 dB increase in sensation level.