Vehicle sound quality has become an important basic performance requirement. Traditionally, automobile noise studies were focused on quietness. It is now necessary for the automobile to be more than quiet. The sound must be pleasing.This paper describes a development process to improve both vehicle noise level and sound quality. Formal experimental design techniques were utilized to quantify various hardware effects. A-weighted sound pressure level, Speech Intelligibility, and Composite Rating of Preference were the three descriptors used to characterize the vehicle's sound quality. Engineering knowledge augmented with graphical and statistical techniques were utilized during data analysis.The individual component contributions to each of the sound quality descriptors were also quantified in this study. This paper discusses the importance of measurement studies to ensure desired experimental precision, the use of regression analysis to overcome the effect of engine rpm during experimentation, and normal probability plots as an initial empirical model building technique.