Mapping perceptual attributes of sound to product design choices

At present, the engineering of product sound is impeded by a lack of clear correspondence between consumer perceptions of product sounds and instrumental measures (physical metrics) of those sounds. This paper discusses the use of “Sensory Profiles” (SP's) in forming a bridge between the component and structural decisions made by product designers, and the perceptual reactions of a user to the product sound. SP's can describe both a target product and the effects of deviations from the ideal on the target profile. In this study, the SP's for different washing machine and vacuum cleaner sounds were determined by applying Principal Components Analysis to a scoring of word descriptors for sounds, obtained using a panel of “expert listeners”. Transformation matrices were formed in order to relate these SP's to consumer ratings of acceptability of these same product sounds, and to various physical metrics which were designed to reflect physical configurations of these products. A sensitivity analysis showed that these transformations or “mappings” were robust to small perturbations in the values of the matrix elements. Since it is possible to relate these physical metrics to the SP's, and it is possible to determine a “best” SP for a product sound, this procedure offers a way to indirectly relate a “good sound” to the metrics. Research supported by the National Science Foundation.