Auralization of road-traffic noise may be a useful tool for city planning, for instance as a support to decisions regarding noise mitigation. However, to be useful, the auralizations need to be perceptually valid. That is, the auralized sounds should be perceptually indistinguishable from real sounds or, at least, similar with respect to perceptual factors crucial for correct decisions. For this reason, the auralization methodology developed in the Swedish LISTEN-project was perceptually evaluated. In four listening experiments, listeners assessed recordings and auralizations of the same car passages. Although real and auralized sounds were not completely indistinguishable, perfect discrimination was not possible. Moreover, good agreements between auralized and real sounds were found for perceived similarity, perceived annoyance and perceived speed. The results illustrate the usefulness of psychoacoustic methods and multivariate statistics for perceptual evaluation of auralizations and provide support for the validity of the LISTEN-approach to auralization.