Affective Assessment of Automobile Sound Quality: Evaluation Framework of AUI (Auditory User-Interface)

Abstract

Automobiles are recognized as more than simple transportation: they are a personal space which provides enjoyment to drivers

as such, various affective elements are embedded within an automobile. Beyond the standardized mechanical performances of automobiles, diverse attempts are undertaken by automobile companies to satisfy affective needs of users. An automobile, a singular product closely tied to the human society, is composed of multiple complex elements. Automakers need to have superior design to satisfy all five human senses. To date, the majority of research on inherent emotions generated by automobiles has focused on visual aspects. However, compound functions of automobiles and multifaceted demands from users demand more research into various affective satisfactions such as auditory and tactile senses. The primary task for humans when driving an automobile is to keep ones eyes forward. Therefore, researchers have concentrated on minimizing sensory disturbances and providing satisfying affective experiences while driving. Previous studies on affective quality regarding sounds of automobiles were limited to identification of affective variables in specific components of sounds and their relationships with sound parameters. This study is driven by a recognition that such limited trends call for an investigation of parameters arousing affective response of drivers along with their sound characteristics. The objective of this study was to construct an evaluation framework and model to predict the manifestation of these affective levels. Since there are limitations upon establishing an affect model for every sound produced from an automobile, representative sounds were selected from various methods in order to investigate the most influential sounds for users and their affective experiences. First, a sound pool was created by collecting all mechanical and electronic sounds from an automobile to derive affective response. These sounds were classified into engine sound, moving/operating sounds, and auditory display sounds based on their characteristics and sound sources. In accordance with these classified sound domains, literature reviews and user survey were conducted to select sounds that played significant roles in auditory affect. Systematic characteristics of classified sounds were identified to develop sound parameters for models of auditory affect. Finally, for engine and moving/operating sounds, modified psycho-acoustic parameters were developed by taking time and Bark scale into account. In the case of auditory display, sound parameters used in automobile as well as aircrafts and medical devices were also considered. Second, the model of auditory affect was designed based on sounds from automobiles. The model of auditory affect adopted a hierarchical structure, just like previously proposed affective models. Model of auditory affect was composed of three dimensions: sensitive, descriptive, and evaluative. Furthermore, each dimension was in hierarchical relation to one another. The model of auditory affect in an automobiles was developed in three stages. First, affective variables were collected by literature reviews and interviews from experts. Opinions of end-users regarding sounds of automobiles obtained from social media were also utilized. Second, these collected affective variables were classified by characteristics of auto-mobile sounds and distributed to each corresponding affective dimension accordingly. Lastly, representative affective variables in each dimension were selected to represent automobile sounds. Third, the model was designed based on sound classifications and affective dimensions proposed in previous studies. However, a new program was developed to evaluate and analyze the model. The model of auditory affect was verified through case studies of three representative sound categories. Results of this study revealed that affective variables in lower dimensions described more than affective variables in higher dimensions. Moreover, the designed sound para-meters were found to be significant in the regression model to predict affective variables of each dimension. This research is expected to aid automobile sound researchers or designers to design new sounds or utilize previous sounds by identifying user affects. Moreover, this study can be applied in the management of user affective quality and its measurement for sounds of specified products by designing the automobile sounds which correspond to classified auditory affective dimensions based on previous models of affective dimensions. The methods to derive elements of affective dimensions proposed in this study can be applied in products other than automobiles as well. In addition, if a database set of recorded sounds and corresponding evaluation results is constructed in managerial perspective, it is possible to conduct the jury test of existing auditory evaluation and investigate the latent needs and affective response of users in more effective manner so that the whole evaluation process can be improved in the future.