People make judgements of the sound produced by vehicles in a variety of situations and contexts. The most common type of assessment is an overall judgement of sound quality like pleasantness and sportiveness. From the car manufacturer''s point of view it is important to note that subjective judgements of sound influence the buyer''s opinion of the performance and value of the vehicle. Consequently, there is a need to quantify the subjective perception of the sound quality associated with a vehicle.This paper presents a study of the sound quality of vehicle exhaust noise based on the correlation between the subjective perception associated with this type of sound and measured psycho-acoustic parameters such as loudness, sharpness and roughness.The tailpipe noise of a selection of representative vehicles was recorded using an artificial head. A subjective evaluation of these sound recordings was made by a jury using standard relative techniques. The data obtained from the jury was then statistically correlated with measured psycho-acoustic parameters using multiple regression analysis. As a result, we obtained a mathematical model of the subjective response that can be used as a measure of sound quality.We finally used the model described above to define the spectral targets associated with a desirable subjective perception for tailpipe noise in terms of pleasantness and sportiveness. This spectral targets are particularly useful for designing and tuning exhaust systems for vehicles.