This article reports a complete two stage hedonic analysis for road noise. For the estimation of the hedonic price function I develop a spatial research design which simultaneously reduces the risk of omitted variable bias and the risk of measurement error in the noise measure. The preference parameters are identified following the approach developed in Bajari and Benkard (2005) by using a simple functional form for utility. Preferences are very heterogeneous and observable demographic characteristics explain 30 percent of the variation in taste for quiet. Results are used to discuss willingness to pay for noise reductions from two policy measures.