Statistical Evaluation of an Annoyance Index for Engine Noise Recordings

The object of this paper is to describe a statistical approach to objectively evaluate the subjective perception of the annoyance caused by engine noises using multiple regression analysis based on subjective data and given physical parameters. About 300 test persons were involved in rating the annoyance of engine noises. 60 representative engine noises, each of them characterized by 48 physical parameters, were studied. This study selects an optimal set of parameters, estimates the regression coefficients so as to minimize the deviation between calculated annoyance scores and real ratings given by test persons and thus reproduces subjective annoyance scores by means of selection and combination of objective physical parameters.