Effects of presentation method and duration on alarm detection threshold in the presence of loud pink noise

Abstract

Detection of pure tone signals in the presence of noise has been thoroughly studied. Most of these studies have used monaural presentation of audio stimuli. Also, studies testing alarm detection in the presence of noise are limited. In 2013, Karunarathne et al., conducted a study and found out that human listeners were able to detect an alarm in negative signal-to-noise ratios (SNRs), as low as −24 dB. This study aims to investigate the effects of presentation method and duration of the alarms on detection threshold. Eight conditions varied by presentation method (monaural vs spatial) and alarm duration were tested. Sixteen human subjects with normal hearing were given the task of identifying which one of two sound intervals contained an alarm along with 80dBA pink noise. Thresholds were estimated as the 79.4% points on the psychometric functions, using adaptive 2-Interval Forced Choice (2IFC) procedure with a 3-down 1-up rule. Results indicated that detection thresholds were statistically significantly lower in spatial condition compared to monaural. The effect of alarm duration was not significant in both spatial and monaural conditions. Thresholds lower than −30 dB SNR were observed in the spatial condition, which agreed with the findings of Karunarathne et al. and further extended threshold boundaries.