Targeting sound quality metrics using sound synthesis

The purpose of jury evaluations in sound quality analysis is to relate customer preference or subjective rating to a measurable quantity in the sound, This is normally achieved by auditioning a range of measured sounds with varying levels of the metric under evaluation to determine a numerical correlation between preference and metric by preference ranking. The problem with this approach is that the target metric may not be the only parameter that varies among the auditioned sounds, At best these other variations reduce the statistical accuracy of the results, at worst they can mean that the evaluation is correlating the wrong metric. The authors propose a technique which focuses the jury evaluation process on a specific metric and show how the technique was applied to the evaluation of powered automotive seat adjusters.