Soundscape assessment with classification of human perception to sound environment

Abstract

To investigate the effect of human sound identification differences on urban soundscape perception, we propose the urban soundscape categorization method based on human attention for sound sources. Virtual reality technology was applied with audio-visual data collected in various urban environments to reproduce realistic environment in laboratory condition with high ecological validity. Fifty subject's response were collected about sound sources identification, perceived affective quality (8 typical attributes and 116 extensive attributes), and overall quality. Human categorization with sound source identification were divided into 3 clusters; cluster 1: attentive to traffic and other noise, cluster 2: less-attentive to sound environment, cluster 3: attentive to natural sound and human-being sound. The results show that soundscape perceptual dimensions were derived differently for each cluster and only the cluster 2 derived new dimension: tranquil and relaxed soundscapes. The soundscape perception model showed that the overall quality assessment according to each cluster are similar in aspect of mean value for each soundscape descriptors, but the indicator which effects on soundscape descriptor are different according to sound source attention. A new perspective perception model can be utilized to understanding urban soundscape perception considering individual characteristics and can be effectively used in design urban planning