https://cloud.blender.org/blog/cycles-turbocharged-how-we-made-rendering-10x-faster
I was reading about motion blur at above article link, and i wondered.
I wondered wouldnt it be faster if motion blur is a compositor thing based upon previous and past frame(s) difrences
Thus a more advanced vector motion blur ? (with some knowhow of pixel movement).
i might be wrong but i think then stills could be rendered and a morph like 2d image processing could blend the blur effect
without depending on some complex calculation for a still image. As essentially the previous or future frame contains info of whats behind the moving object and thus basedupon that a blur factor with background could be determined.
Just curious, i might be wrong but i wondered this