Automating Urban Soundscape Enhancements with AI: In-situ Assessment of Quality and Restorativeness in Traffic-Exposed Residential Areas

TL;DR

An automatic masker selection system (AMSS) utilizing natural sounds to mask (or augment) traffic soundscapes is implemented, affirming the efficacy of AMSS as a soundscape intervention, while streamlining the labour-intensive assessment of"Pleasantness" with probabilistic AI prediction.

Abstract

Formalized in ISO 12913, the "soundscape" approach is a paradigmatic shift towards perception-based urban sound management, aiming to alleviate the substantial socioeconomic costs of noise pollution to advance the United Nations Sustainable Development Goals. Focusing on traffic-exposed outdoor residential sites, we implemented an automatic masker selection system (AMSS) utilizing natural sounds to mask (or augment) traffic soundscapes. We employed a pre-trained AI model to automatically select the optimal masker and adjust its playback level, adapting to changes over time in the ambient environment to maximize "Pleasantness", a perceptual dimension of soundscape quality in ISO 12913. Our validation study involving ($N=68$) residents revealed a significant 14.6 % enhancement in "Pleasantness" after intervention, correlating with increased restorativeness and positive affect. Perceptual enhancements at the traffic-exposed site matched those at a quieter control site with 6 dB(A) lower $L_\text{A,eq}$ and road traffic noise dominance, affirming the efficacy of AMSS as a soundscape intervention, while streamlining the labour-intensive assessment of "Pleasantness" with probabilistic AI prediction.

Figures (5)

• Figure 1: Study sites in a public residential estate in Singapore: (a) A ground-floor pavilion (GND) in the outdoor recreational area at coordinates (1.401358, 103.895427). (b) A rooftop garden pavilion (ROOF) situated atop an 8-storey multi-storey car park at GPS coordinates (1.343373, 103.686134). (c) An overview of the end-to-end process of the automatic masker selection system (AMSS)
• Figure 2: Overview of experimental procedure and data collected from participants for the in-situ validation experiment.
• Figure 3: Simple contrast of means across all perceptual attributes organized by condition and site. Contrasts by condition are between group at each site, whereas contrasts by site are within group for each condition. The scales for all attributes are normalised to the range [$-$1,1]. Significant differences as determined by posthoc contrast tests are accentuated
• Figure 4: Energetic mean A-weighted, fast time-weighted sound pressure level, $L_{AF}$, of the loudest binarual channel across all sessions in the AMSS and AMB groups at ROOF without aircraft flyby. The shaded error envelope represents the standard error of the mean.
• Figure 5: Mean perceptual ISOPL, OSQ, PRSSFas, PRSSBA, and PRSSCom scores across all participants per session (y-axis) as a function of normalized objective $L_\text{A,eq}$, $L_\text{C,eq}$, and $N_\text{95}$ scores of each session (x-axis). Fifty percent of the sessions lie within the median contours computed for AMB--GND, AMB--ROOF, AMSS--GND, AMSS--ROOF contrast subgroups. The left to right columns represent $L_\text{A,eq}$, $L_\text{C,eq}$, and $N_\text{95}$, and each row represent each of the perceptual metrics, respectively.