Prediction model of soundscape assessment based on semantic and acoustic/psychoacoustic factors

Abstract

While conventional sound/noise environment assessment, management and relevant standards are based on the objective measurement of sound pressure level (SPL), soundscape research showed that humans’ subjective assessment (such as positive and negative) cannot be well explained by the SPL or other acoustic indicators alone. Rather, soundscape assessment depends on many more factors such as cognition and context, which also need to be considered as suggested in the new soundscape standard. Among the different factors, a large number of previous studies supported that semantic factors, i.e. the sound sources which compose the sound environment—e.g., natural sounds (moving water, bird song, etc.), mechanical sounds (road traffic, construction, etc.) and human activity sounds, play the most important role in human perception and assessment. Therefore, from the systematic literature review of the relationship between such semantic factors and soundscape assessment, this present study proposes a model based on the sound source information, as well as SPL and psychoacoustic indicators of each of the different sound sources, to explain/predict largely, even though not fully, the soundscape assessment. Furthermore, for the practical use of this model which would be independent of manual efforts in recognition of sound sources through listening, this study discusses the possibility of integrating a computer algorithm of sound source recognition from field recordings of sound environment, for automated soundscape assessment.