Search

In a recent paper [Acta Acustica united with Acustica 93 (2007) 73―83] [1], an experimental noise annoyance meter was proposed for estimating acceptable noise levels, L Aeq7 . In this experiment, the subjects were able to relate the value of L AeqT to the differences between the seven-point annoyance scale scores displayed. Therefore, they directly gave the value of L AeqT for their acceptable threshold. This study investigated whether each acceptable L Ae q T level can be predicted from the annoyance and noise sensitivity. First, multiple regression models for each outdoor noise source were constructed to predict the annoyance from L Ae q T and the noise sensitivity based on the data collected in and around the city of Nagano. The regression models were used to determine whether each outdoor noise was over or under the acceptable threshold, which was defined as the midpoint between the scores "a little annoying" and "somewhat annoying." The correlation between the experimental results using a separate sample of residents from other prefectural capital cities in Japan and the predictions by the regression models was 0.703. Also, the discriminant models were used to classify cases by their values of a categorical dependent, where correlation with the proposed models was 0.837, indicating that the discriminant models are valid.