This question likely comes from a lack of understanding of the technology, but could this same mechanic be used as a means of super sampling/upscaling?
For example, say you used a GOUT frame as a source and a 70mm frame as a reference (both would be the same frame of course). Could you theoretically use an algorithm to create an upscaling script that would match the detail and idiosyncracies of the higher quality source as best as possible?
I’ve always wondered this. In theory this could be the best upscaling software out there.