Annoyance modeling of construction noise using acoustical features, noise sensitivity and health condition

The construction site noise is the biggest cause of the noise dispute in Korea. A listening experiment was conducted to produce annoyance prediction models for four kinds of construction noise that cause the most complaints. These noise were generated by the construction machinery such as the pile driver, the excavator, the pumping vehicle, and the concrete mold removal working. These four kinds of noise recorded at the construction site were adjusted to the stimuli with 35 ~ 80 dB(A), and they were played to the subjects and the annoyance score was surveyed. The objective indicators of the stimuli and the survey results of subjects were used as variables to develop the annoyance models caused by construction noise. The annoyance models were developed by multiple regression. Acoustic features include traditional indicators such as Leq, Lmax, as well as indicators of spectral features and temporal features proposed in recent studies. And sound quality indices were used to model. These were calculated mean, maximum, and percentiles values of loudness, sharpness, Roughness and fluctuation strength. Also noise sensitivity and health condition were used as the variables of annoyance prediction models. These were obtained from subjective questionnaire.