I've currently been getting into shaders and done a handful of effect for c3.
When reading resources about shaders I often find the technique of down-sampling and then running shaders on that downsampled framebuffer and then upsample later. This seems to lead to crazy performance benefits as the amount of pixels is just that much lower.
(example, super fast blur Dual Kawase https://www.shadertoy.com/view/3td3W8)
I tried to apply similar techniques in c3 but that either lead to very weird results or seem outright impossible. So I wondered is this a skill issue and it's possible?
Or if not possible I wonder if it would be feasible to extend the effects sdk to allow to read and write to frame buffers to allow for some additional shader techniques.