Crowd noise simulation system for sound environment evaluation of public spaces

Abstract

Aiming at the prediction and evaluation of sound environment of public spaces, a simulation-based auralization system is developed for reproducing background crowd noise. The system combines room acoustics modeling using a ray tracing method with a six-channel sound field reproduction system in an anechoic room. As sound sources, footsteps and voices of pedestrians moving on a floor, and HVAC noise from a ceiling are modeled. In a preliminary examination simulating several existing public spaces, a general correspondence to real sound field are confirmed in noise level and auditory impressions such as noisiness, liveness, and so on. Next, simulating a variety of imaginary rooms with changing room dimensions, absorption and pedestrian density, a subjective experiment of auditory pairwise comparisons is carried out. The results show that the noisiness fairly corresponds with the noise level, whereas the feeling of reverberation does not clearly related to the reverberation time, which may be due to the continuous overlap of sounds.