Characterizing impulsive noise with A-weighted sound pressure, kurtosis, and higher order statistical moments

Permanent hearing loss and tinnitus are the first and second most common disabilities for current members and veterans of the US armed services. The recently finalized Military Standard 1474E for noise uses the peak sound pressure level, A-weighted equivalent level and the Auditory Hazard Assessment Algorithm for Humans (AHAAH)model. An alternate approach uses the kurtosis statistic to indicate the extent of impulsiveness of the sound pressure waveform. This paper discusses how the characterization of impulsive noise by the A-weighted equivalent level can be supplemented with kurtosis. Impulsive noise is processed in the time domain using A-weighting filters and one-third octave band filters. Each one-third octave band has its own analysis time window that is the maximum of five wave periods or 0.01 seconds. For all one-third octave bands, the sound pressure waveform is processed using a time step of 0.01 seconds. Overlapping of the analysis time windows may occur below 500 Hz. The short time step provides a higher temporal resolution for modelling the acoustic reflex. A dataset of fifty previously collected acoustic waveform stimuli are used to test the signal processing methods. The dataset was originally collected by SUNY-Plattsburg, New York and USAARL in Fort Rucker, AL.