卧室声景体验要素及理想声景营建研究

发布时间：2018-04-05 04:37

本文选题：卧室声景　切入点：支持向量机　出处：《哈尔滨工业大学》2017年硕士论文

【摘要】：卧室是人类最重要的生活空间之一,人每天平均会有三分之一的时间在卧室中度过,不好的卧室声音环境会干扰人们的日常生活,甚至对人的身心造成损害。随着时代的发展,人们对声音环境的要求逐渐提升。声音环境设计的目的不再是单纯的屏蔽噪音,而是提供更加符合个性化的需求的声音环境服务。声音景观的主要思想是将环境中所有的声音、社会因素及文化因素纳入研究,引入视觉与听觉环境相协调的概念,以达到更好地满足人们对声音环境的需求的目的。本论文将针对卧室,开展实地调研,获取声景主观评价和声景因素数据,并建立了实地调研卧室的声场仿真模型,对影响卧室声景影响因素和主观评价进行了多元分析,并根据分析结果建立卧室声景主观评价多类支持向量机模型,以该模型为参考,寻求满足人们需求的理想卧室声景的营建方法,促进了以人为本的设计理念。首先,收集资料获得的知识及理论,选取调查样本和实地测试地点,进行访谈,设计卧室声景要素调查问卷。对测试地点进行实地调研,包括问卷调查和声音信号的采集两部分,通过实地调研可获得卧室声景研究数据。据通过对声音信号进行分析,可获得卧室声环境心理声学参数等声学数据。而后续展开的研究是建立在实地调研所得数据的基础之上的。其次,对声音采集卧室的空间结构、家居家装材料、声源位置和温度湿度进行记录,依据记录使用声学软件建立卧室声环境仿真模型,并对3处实测卧室声场进行仿真,将获得的声场仿真结果与实地采集数据进行一一比对检验,选择适当的声场计算公式。对所有调研卧室进行声场仿真,并计算得混响时间,并将其作为卧室声景数据之一。最后,以问卷及声学分析所得声景因素数据作为研究样本,通过因子分析挖掘卧室声元素深层关系,提取新的声景因素。进行多元统计分析,得到卧室声景要素;再以卧室声景要素数据作为研究样本,运用线性回归、多类支持向量机和网格搜索等算法建立卧室声景评价模型并进行验证,并对卧室声景营建工作提出建议。该模型可根据声景要素对用户声景评价进行预测,为理想声景营建工作提供参考,使卧室声音设计更具有针对性。
[Abstract]:The bedroom is one of the most important human living space, people there will be a daily average of 1/3 of the time spent in the bedroom, the bedroom is not good sound environment will affect people's daily life, even on people's physical and mental damage. With the development of the times, people gradually improve the requirements for sound environment. The purpose of the design is no longer sound environment is a simple noise shielding, but to provide sound environmental services more in line with the needs of individual. The main idea of sound landscape is all the sound environment, social and cultural factors into the study of factors, introducing visual concept of harmony with the auditory environment, to better meet the needs of the people on the sound environment for the purpose. This paper will focus on the bedroom, to carry out field research, data acquisition of subjective evaluation of soundscape and landscape factors, and establishes the simulation model of acoustic field research on the bedroom, lying Multivariate analysis of factors affecting room soundscape and subjective evaluation, and according to the analysis of the establishment of the subjective evaluation of soundscape bedroom multi class support vector machine model, with the model as reference, to meet the construction needs of people's ideal bedroom soundscape, promote the people-oriented design concept. Firstly, knowledge and theory of data collection the survey sample selection and field test sites, interviews, bedroom design soundscape elements questionnaire. The on-the-spot investigation to the test sites, including the questionnaire and voice signal acquisition part two, through the investigation of available data. According to the bedroom soundscapehas based on sound signal analysis, can obtain the bedroom environment psychoacoustic parameters of acoustic data. The following research is based on the data from field research. Secondly, the spatial structure of the sound collecting bedroom, Home Furnishing decoration materials, the location of the sound source and the temperature and humidity were recorded on the basis of the record established bedroom environment simulation model using acoustic acoustic simulation software, and the 3 bedroom at the measured sound field, the sound field simulation results and field data are compared one by one test, choose the suitable calculation formula of sound field. The acoustic field simulation of all research the bedroom, and calculate the reverberation time, and as one of the bedroom soundscape data. Finally, the data analysis of the questionnaire and the acoustic soundscape factors as research samples, through factor analysis of acoustic element deep relationship between mining bedroom, extraction of soundscape new factors. By multivariate statistical analysis, obtained the soundscape elements to the bedroom; in the bedroom of soundscape elements as the research sample data, using linear regression, multi class support vector machine and grid search algorithm to establish your soundscape evaluation model and verification, and to lie Suggestions are made for the construction of room soundscape. This model can predict users' soundscape based on soundscape factors, provide reference for ideal soundscape construction, and make bedroom sound design more targeted.

【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU241;TU112

【参考文献】