Predicting overall reaction to multiple noise sources

Abstract

The validity of models predicting overall human response to complex sound environments, which are constituted of road traffic and railway noises, is discussed using results of psychological experiments carried out both in a laboratory and outdoors, as well as by using results of social surveys. Existing predictive models (energy summation model, independent effects model, response summation model, and so on) were applied to the data. The result shows that the simple energy summation model gives the poorest predicting ability in every case. It suggests that differences of response among different noise sources must be considered. The effective level D of road traffic noise relative to railway noise was calculated using Ollerhead’s response summation model as approximately 10 dB in the laboratory experiment. The D values decreased as compared to the above value in cases of field experiments and social surveys, but it was still clear that the contribution of road traffic noise to the overall judgment was greater than that of railway noise. Judging from the coefficient of determination and standard errors shown in the results, the independent effects model seems to be the most suitable to predict the overall human response to complex sound environments.