The prediction of sound levels in vehicles and the environment is now a common practice in automotive engineering. This has been made possible by the development of powerful tools based on boundary element techniques. These “first generation” systems provided the user for the first time with easy to use tools to predict sound levels which result from any known set of vibrations of the panels surrounding the acoustic region. This paper will describe a “second generation” acoustic system based on boundary elements, which not only enhances the accuracy of the modeling, but also provides powerful diagnostic facilities to enable the user to identify the main contributions to the sound intensity at any point of interest.Applications are presented, and include a classical automotive engineering problem in which the software system identifies the body panels which contribute most to the noise level at the driver's right ear. Accurate information of this type is important in improving the quality of the ride experienced by the driver, as panels which are causing noise problems may be remedied by stiffening or by covering with different sound absorbing materials.