Work

Greenery, sky and water in view and how people judge urban soundscapes

Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Street-level imagery is widely used to measure how urban public space looks, and increasingly how it sounds. It measures visual scene composition, namely how much of a view greenery, sky, buildings and other categories occupy. Whether composition captures the visual influences on soundscape evaluation is rarely tested at the same moment. The study analysed 2,262 on-site soundscape questionnaires from 12 urban open public spaces in Shenyang, a northern Chinese city. Each had segmentation and image metrics of the respondent's view and, for 2,229, a binaural recording. Three findings emerged. First, composition explained and predicted almost none of overall evaluation, pleasantness or harmony beyond perceived sound sources and loudness, which explained about thirty times as much variance. This held out of sample and at held-out sites. Eventfulness showed a small predictive gain within sites, associated with visible vehicles, but this gain did not extend to held-out sites. Second, scene condition, the appearance of what is seen, revealed an association that composition missed: greater sky blueness was associated with higher overall evaluation beyond the auditory measures, whereas the amount of sky was not. Greenery share was constant between autumn and spring while its colour content fell by a fifth. Third, visual associations varied with perceived sound characteristics, perceived loudness and the sound sources heard. Visible water helped more as perceived loudness increased, sky blueness did not, and traffic both heard and seen lowered evaluation beyond its separate effects. Based on these results, scene composition has limited value in explaining or predicting soundscape evaluation. Visual assessment and design should instead focus on sky condition and on specific elements such as water in loud settings.
Language: English
Open Access status: green
OpenAlex ID: W7237212879
Cites: 51
Cited by:
Related to: 10