In this report we benchmark the plane-to-plane objective quality metric. This is, a metric that measures the angular similarity of tangent planes between two point cloud models and relies on normal vectors that are carried with associated pairs of points. Evidently, the performance of the metric depends by default on the normal vectors and the way they approximate the underlying surfaces. Thus, to shed some light on the impact of the normal estimation algorithm selection and configuration in the performance of the metric, we choose three widely-used normal estimation algorithms, and we test several neighborhood sizes over which the normal vector of a point is estimated. Then, in a first step we evaluate the normal estimation error and, in a second step, we benchmark the plane-to-plane metric against two subjectively annotated datasets, as a function of the selected normal estimation algorithms and their configurations.