Linking ecoacoustic indices to psychoacoustic perception of the urban acoustic environment

Research about human perception of the acoustic environment is dominated by the use of sound pressure levels (SPL) or proprietary psychoacoustic indicators, limiting the link between sound and perception to only these two approaches. The aim of this study is to make connections between soundscape perception descriptors and ecoacoustic indices so that a perceptual dimension of ecoacoustic indices is ascertained. The study employs a laboratory experiment where 309 participants were exposed to sound and imagery from nine different urban land uses (n = 2,781) according to the soundscape protocol DIN ISO 12913-2. Spearman's correlation and ordinal logistic regression are used to analyze the relationship between continuous numeric ecoacoustic indices and Likert scale psychoacoustic perception outcomes. We find that the predictors median amplitude (M), acoustic richness (AR), number of peaks (NP), and acoustic evenness index (AEI) have significant correlations (r > 0.5: p ≤ 0.05) and effect sizes (β ≥ 0.1) for the psychoacoustic outcomes perceived affective quality (PAQ) of calm, pleasant, chaotic, annoying, and soundscape identifiers (SSI) traffic sounds and natural sounds. We find that between 59.4 % and 81.9 % of the variation within dependent SSI and PAQ variables is explained by ecoacoustic indices based on Nagelkerke's R-Squared values. High M and AR values are indicators for sound environments with high median or consistent amplitude, such as traffic noise, and high NP and AEI values are indicators for high fidelity sound environments with natural sounds or human beings that may be perceived as calm or pleasant. We conclude that combinations of M, AR, NP, and AEI as composite indicators contribute to understanding nine of the 14 psychoacoustic perception outcomes in DIN ISO 12913-2.