Most video quality metrics (PSNR, VMAF) were built for compression in filmed video, so they miss a lot of what goes wrong with temporal upscaling and anti-aliasing in games: blur, ghosting, flickering, aliasing, moiré patterns, fireflies, and shimmering
Karol Cyganik (AMD) and Dariusz Brzeziński (Poznan University of Technology) built Video Rater, an open-source framework for collecting human ratings of game engine output side by side with ground truth. The goal is to gather the kind of data needed to build better metrics for rendering artifacts.
Read the paper on GPUOpen: