Noise measurements made for the purpose of hearing conservation practice have as their objective the extraction of some physical metric from the noise that can be used to estimate the hazards to hearing from prolonged exposure. Current practice continues to rely primarily upon a time averaged, integrated approach to quantifying a noise (i.e., an energy metric) despite substantial evidence of the inadequacy of this measure in a variety of industrial situations. Recent experiments have shown that conventional measures of noise exposure such as Leq (equivalent level) and spectrum were poorly correlated with trauma, but that metrics that quantify the temporal characteristics of the noise such as frequency domain kurtosis were highly correlated with both the frequency specificity of hearing loss and the severity of trauma. The primary objective of the proposed research is to develop analytical methods based upon contemporary signal analysis technology that can be used to better characterize noise exposure for the purpose of hearing conservation. The wavelet transformation and the bicepstrum will be developed to obtain metrics which will quantify the temporal characteristics of an exposure and which when used in conjunction with conventional measures will more accurately assess the potential of that noise for producing hearing loss. Specifically, the distribution of time domain kurtosis, the joint peak- interval histogram and the frequency domain kurtosis will be the primary metrics that will be obtained from these analytical techniques.
