Vehicle interior noise consists of a superposition of broadband contributions from powertrain, wind, and tire-road noise. Tire-road noise has become increasingly important referring to overall acoustic comfort, especially for (luxury) sedans with pleasant low-noise engine sounds. An interior noise recording during a coast-down (engine switched off) contains different components: a mixture of wind along with airborne and structure-borne tire-road noise shares. Separating the mixture into these components requires appropriate algorithms and additional measurements. Therefore, structure-borne excitation signals as well as the airborne noise radiation of all four tires are measured simultaneously to an artificial head recording in the vehicle interior during a coast-down test from maximum vehicle speed to standstill. Based on the fact that wind and tire-road noise are uncorrelated, evaluating the multiple coherence between the excitation signals and the simultaneous binaural recording allows calculating speed-dependent FIR filters to separate the different components. In this paper a new approach is presented using Operational Path Analysis (OPA) to estimate the transfer functions of a Multiple-Input-Multiple-Output (MIMO) model of the tire-road noise contributions based on road measurements. An important goal is the high quality auralization of the overall sound and the contributions of the different sources, for example the interior noise share caused by a single tire or even the corresponding airborne and structure-borne contributions. The uncorrelated wind noise can be determined as the difference signal between interior noise and synthesized tire-road noise. In the case of uncorrelated excitation signals OPA is very efficient and accurate; no additional laborious transfer function measurements are required. The advantages and possible drawbacks of OPA for characterizing wind as well as airborne and structure-borne tire-road noise in comparison to coherence filtering will be discussed.