been pushing for realtime afterburn for a while… we’ll see. on one level im dissapointed how well baked flipbooks get the job done. incredibly inexpensive and the human eye doesn’t seem to care too much that the lighting makes no sense. which makes me sad.
I’d made some progress making these “volume boxes”, need to revisit it. ended up failing at inversing the transform for the raycast so that you could spin, stretch, and scale it just like normal particles. need to revisit that.
what you’ve got to ask yourself if your looking ahead is whats going to make significant bounds next gen. bandwidth or alu? My hunch is abusing the texture bandwidth is going to make the most sense. We’ll be flush with memory on the card. So baking as much as you can is going to probably be possible as 8k everything will seem normal. I’m trying to look into a way of baking better imposters. But if ALU somehow makes great leaps and you can start doing all these 3d noises efficiently then raycasting through them is perfect as it doesn’t need to call a single texture, which just like auto stalls the gpu.
personally the next 3 years seem more about honing the artistic craft using the current gen set of techniques, as the hardware is fixed, and most people are dev-ing atleast partially for consoles.
the excitement of new hardware was super cool though! Soon it’ll be back!