There is a need to develop practical standard sound-power tests for manufactured products independent of the test location. That it is possible to develop such standards, for use in a normal work area, has been demonstrated in tests with six light vehicles. The sound power of each vehicle was measured for different engine speeds and loads using a semicircular array of sound-intensity probes rotated over a hypothetical, hemispherical surface enclosing the vehicle running on a dynamometer roll. The presence of the walls and ceiling of the work area appears to have a negligible effect on A-weighted sound-power measurement. Good agreement was obtained with the overall sound power of a standard reference source. Results for the six vehicles indicate that the peaks in the narrow-band sound-power spectra occur at frequencies corresponding to the orders of excitation of components of the vehicle power train. This has diagnostic value because it identifies noisy components and provides a quantitive measure of their contribution to the total sound power of the vehicle. It could also be used to detect malfunctioning and wear of components.