What if we randomly sampled points?
Is there some heuristic for how many samples would be enough to represent the signal to some level of accuracy, or do we have to adjust it repeatedly until the result seems good?
Do we think of this as real-time sampling (in the case of sound for example)? If we don't sample a point in time then it's lost to us, so we need to come up with a sampling method beforehand?
will the way of sampling (random sampling, sampling by specific numbers, pixel ratios) be possible approaches
Does this concept of supersampling in graphics also depend a lot on the computing power of the GPU? I don't know much about graphics but I was curious if this is a factor.