The color of automotive coatings varies with illumination and detection angles. For 3D rendering of these coatings currently two categories of methods exist, aiming at either color accuracy or at computational speed. Current methods that aim for color accuracy are based on measurement data from BRDF instruments. Since many hundreds of measurement geometries are needed to capture the color variation with angles, these methods require timeexpensive interpolation of large three-dimensional Look-Up Tables. Current methods that aim for computational speed use physically crude approximations, usually taking into account color variations with respect to only one of the four angular dimensions. Here, we derive a new approach for 3D rendering of automotive and other gonio-apparent coatings, which is a dedicated form of microfacet models. It aims at improved color accuracy as compared to the current computationally inexpensive methods, combined with higher computational speed and lower cost as compared to current color-accurate rendering techniques. The new approach utilizes a recently developed physical analysis method, introducing flake-based parameters and isochromatic lines, for the reflection properties of automotive coatings. This makes it more accurate than current fast rendering methods. The new method naturally leads to two-rather than three-dimensional Look Up Tables, which explains the small computation time it needs. We show that when applied to 3D rendering, this method indeed leads to accurate 3D rendering of automotive coatings while requiring reduced computation times. For numerical errors found in some special cases, solutions are found and tested.