Could this be used for training AI to do the reverse, i.e., generate high-quality videos from older analog TV and VHS artifacts? Start from high quality videos, use this or a similar library to generate analog/VHS videos that now have the ground truth available, train AI from the data.
Reddit - ukrainewarvideoreports - the Ukrainians are taking their relatively poor 'Last frame' drone shots and ai-ing them to be shockingly-detailed images of people about to die.