Multilevel Annoyance Modelling of Short Environmental Sound Recordings

The recent development and deployment of Wireless Acoustic Sensor Networks (WASN) present new ways to address urban acoustic challenges in a smart city context. A focus on improving quality life forms the core smart-city design paradigms cannot be limited simply measuring objective environmental factors, but should also consider perceptual, psychological health impacts citizens. This study therefore makes use short (1–2.7 s) recordings sourced from WASN Milan which were grouped into various sound source types given an annoyance rating via online survey with N=100 participants. multilevel psychoacoustic model was found achieve overall R2=0.64 incorporates Sharpness as fixed effect regardless type Roughness, Impulsiveness Tonality random effects whose coefficients vary depending source. These results promising step toward implementing on-sensor features type, is ultimately not dependent level.