Statistical structures of indoor traffic noise in a high-rise city

Noise measurements were conducted in seven residential flats close to the traffic in Hong Kong. Noise-level time variations and their statistics are discussed. The results of the overall noise level statistics reveal that Gaussian noise-level distribution can only be found under the free traffic- flow condition while some characteristics of gamma noise-level distributions are observed when the flow is of the interrupted type. Short-duration noise- level time variation statistics reveal a well-defined relationship between skewness and kurtosis for each traffic-flow pattern identified. Results also suggest that the Pearson type I and IV distributions are useful for describing the short-duration noise-level distributions.