Parameter extraction of traffic noise by analysis of auto-correlation function and interaural cross-correlation function

Analysis on auto-correlation function and interaural cross-correlation function of road traffic noise and railway noise recorded by using dummy head on the spot was carried out. 3 representative parameters for railway noise and 4 representative parameters for road traffic noise were obtained through correlation analysis, principal component analysis and subjective evaluation experiment. It was shown that comparing with traditional noise level measurement, 7 factors that mean the temporal factors extracted from autocorrelation function analysis and spatial factors extracted from interaual cross-correlation function analysis could represent the acoustical information of traffic noise more adequately and accurately. In the process of measurement and analysis on road traffic noise, if the physical factors relevant to subjective diffuseness of noise field and pitch, interaural time delay for the horizontal angle, and energy represented at the origin of the delay were obtained, the basic signal information of road traffic noise that was consistent with subjective noise annoyance evaluation could be grasped thoroughly. The same for railway noise, its basic signal information consistent with subjective noise evaluation could be clarified clearly by getting the physical factors relevant to subjective diffuseness of noise field, interaural time delay for the horizontal angle and effective duration of the envelope of the normalized autocorrelation. Comparing with temporal factors, spatial factors such as interaural time delay for the horizontal angle and the physical factor relevant to subjective diffuseness of noise field derived from interaural cross-correlation function had more effects on subjective noise annoyance evaluation.