Auditory Alarms Design Tool: Spectral Masking Estimation Based on a Psychoacoustic Model

Abstract

Human ability to detect auditory alarms in the presence of noise has been identified as an issue in various working environments, with potentially serious consequences. Spectral masking of alarms is recognized as a contributing factor to response failures. In this paper, a GNU OCTAVE code implementation for the estimation of spectral masking is detailed. The method is based on a psychoacoustic model of the peripheral human auditory system and is suggested as a tool to support the design of efficient auditory alarms. Three scenarios are investigated using standardized clinical auditory alarms as test stimuli: (1) pair of same priority alarms; (2) pair of different priority alarms and (3) alarm in the presence of environmental noise. The implemented method offers a visualization of estimated masked signal spectral components, enabling the sound designer to evaluate manipulations for masking avoidance.