Study On Subjective Evaluation And Analysis Methods Of Vehicle Sound Quality

Sound quality is applied originally in the field of acoustic characteristic evaluation of machinery and electrical equipments, and afterward in the field of automotive and electrical appliances. Now it extends to many fields, such as aviation, railway, traffic noise and environment noise. Research on sound quality has attracted more experts'attentions, and is becoming an international academic hotspot. Domestic research of this field is still at the beginning stage, so acousticians shoulder heavy responsibilities.In the dissertation, the definition of vehicle sound quality is expatiated as well as its research contents, progress and psychoacoustics characteristic. Besides, it introduces main psychoacoustics parameters and mathematics models. Experiment methods, implementation details, and analysis methods of subjective evaluation about sound quality are summarized in detail.Combined with practical work and existing conditions, taking the interior noise samples from a domestic car at constant driving condition as the evaluation objects, a preliminary study on subjective evaluation and analysis methods of vehicle sound quality is carried out with paired comparison method and semantic differential method. By means of questionnaire main sound quality characteristics of samples are found out, and evaluation vocabulary is determined. It is also found that roughness and fluctuation are hard to understand for evaluation subjects. So it is necessary to establish evaluation vocabulary of sound quality used in China. While processing the data of semantic differential method, geometric mean method and percentage conversion method are used and have been proved to be conformable. It is discovered that experienced groups are not always superior to non-experienced groups. Regression equations of annoyance degree and roughness of PC and annoyance degree and fluctuation of SD have been established by correlation analysis and regression analysis. Research on samples from 7 to 12 indicates that PC is more difficult to make a judgment than SD.