Soundscape Classification Of Large Urban Park

As a significant part of the urban landscape and living environment,urban parks provide both important ecological benefits and humanistic vitality for the city.They influence citizens not merely through visual environment but also through the soundscape.In fact,soundscape exists through human perception,and shapes the place as a part of the local socio-cultural backgrounds and context at the meantime.However,soundscape is still a new concept.Former studies have concerned much about the epidemiological aspects of“noise”which belongs to the acoustic environment.In fact,the International Standardization Organization promulgated its definition only for years ago,and the publications in this field are relatively not enough to examine its potential.Overall,more exploration is needed to give specific evidence,provide predictive models,and eventually involve soundscape into the broader framework of urban design as well.In this study,we define the scope of soundscape as the new standard published by the International Standardization Organization in 2014,where soundscape is defined as as a perceptual construct rather than a physical auditory phenomenon.So soundscape is measured and assessed with regard to human perception including physical and perceptual data.The survey is based on the human perception in eleven open spaces of the large urban park.At each site,subjective data are collected with questionnaire in semantic scales,while characterize acoustic environment with acoustical parameters.This dissertation sought to determine principal descriptors and indicators that characterize the soundscape and to classify different acoustic environments of large urban parks with perceptual features and acoustical parameters collected at various sites of Xuanwuhu Park in Nanjing,China.This work mainly consists of three parts:The first part introduces literature review and theoretical induction,where we defines the objective,goals and significances,summarizes the state-of-the-art of soundscape in various subjects and current quantitative methods of soundscape measurement,classification and expression.The second part is field investigation,taking Xuanwuhu Park as the research sample,the combination of records measuring and questionnaire is implemented to collect respectively subjective and objective soundscape data of eleven sites in the park.The third part focuses on discussion and conclusion.Subjective evaluation of soundscape literature with semantic scales was analyzed through Principal Component Analysis,and‘Pleasantness',‘Eventfulness',‘Spatiality'and‘Dynamic'are extracted as four principal components and account for overall 67.94%of the perceived soundscapes in large urban Park.Based on public perception scores from the four crucial factors above,the eleven acoustic environments were classified into three Clusters.To classify soundscapes correctly with hierarchical cluster analysis,two measures were applied to determine the metric of distance between variables.With results compared,the Pearson correlation coefficient is recognized as a better measure to eliminate individual deviation when interviewees are selected randomly from the very location.Also,discriminant function analysis was conducted to determine the physical characteristics of each Cluster,and to provide predictable functions as well as attest the accuracy of the classification results.It is revealed that sound pressure level(L_Aeq),the variability over time(L_A10–L_A90)and spectral contents of sounds(L_Ceq–L_Aeq and log G)could identify the three classified soundscape effectively with overall accuracy more than 81.8%,though it may lack stability when determining those acoustic environments dominated by traffic or construction noise.These three Clusters classified by HCA are well explained by the dominant sound source,perceptual principle factors,as well as acoustic parameters.Cluster A refers to tranquil acoustic environment dominated by natural sounds,and is characterized by the highest value of Pleasantness whereas it is at the lowest level both in sound strength(L_Aeq)and spectral centroid?log G?.Cluster B contains eventful soundscapes with a high percentage of sound sources from diverse human activities.This soundscape shows the highest positive value in Eventfulness,and could be identified when a high but changeful equivalent sound pressure level(L_Aeq and L_A10–L_A90)and the proportion of high-frequency sounds(L_Ceq–L_Aeq)exist at the same time.Cluster C,whose acoustic environment are dominated by noise from traffic and construction,obtains the lowest value both in Pleasantness and Eventfulness,and is characterized by loudness and durative spectral features,although the spectral centroid of traffic noise is incongruent with that of construction.In conclusion,typical soundscapes in Xuanwuhu Park are classified and characterized in both physical and psychological ways through this dissertation.Based on the local condition in Nanjing,these findings provide a valuable reference for designing better-perceived soundscape in large urban parks.