While electric and hybrid vehicles are becoming increasingly common, the issue of engine noise is becoming less important, because it does not dominate the overall noise perceived in the passenger compartment in such vehicles anymore. However, at the same time, other sound sources such as air conditioning, start to emerge, which can also cause annoyance. The CEVAS project, involving VALEO, CETIM, University of Technology of Compiègne, ESI GROUP and GENESIS, deals with the acoustic simulation and perception of automotive air-conditioning (HVAC) and electric battery cooling (BTM) systems. While the other partners focused their work on the aeroacoustic characterization, modeling and simulation, GENESIS’ part in the project is dedicated to HVAC sound synthesis and perception. In order to do the synthesis of the acoustic spectra provided by the partners of the project, an additive model was used. Its ability to reliably reproduce sound recordings was addressed through a listening experiment, which also helped in defining the necessary resolution of the input spectra in order to produce convincing sounds. Finally, various HVAC sounds, including different models, operating modes and airflow rates, were assessed by means of a psychoacoustic method involving two listening experiments: a verbalization task, and a semantic differential task. The results were statistically analyzed and a robust sound quality model based on loudness and tonality metrics was proposed.