Quantification and characterization of the role of subjective preferences on vehicle acceleration sound quality

The sound quality of vehicle acceleration, which directly reflects a consumer's subjective perception of the driving ability and riding comfort of a particular vehicle, has received significant interest. To investigate consumer perception of acceleration sounds, a subjective evaluation with 14 acceleration sound samples was conducted using a grouped pairwise comparison (GPC). After clustering the valid evaluation data, two main subjective acceleration preferences (SAPs) were found. Because a subjective evaluation includes certain misjudgments, a fuzzy reconstructed algorithm is proposed to quantify the subjective evaluation results while a GPC is used. Nine objective parameters highly correlated with the SAP, including psychoacoustic parameters and typical acceleration order characteristics, were selected through a correlation analysis. A support vector regression algorithm with particle swarm optimization (SVR-PSO) was applied for modeling the SAP. Through a quantitative analysis of the relationship between the SAP and objective parameters, two main SAPs, classified as luxury and sporty sound qualities, were characterized. To verify the characteristics of luxury and sporty sound qualities, some acceleration sound samples were modified and compared with the original acceleration sound samples. The experimental results validate the accuracy of the predicted model and the characteristics of the two types of preferred acceleration sound quality. (C) 2019 Elsevier Ltd. All rights reserved.