In the Showdown demo there is an ExplosionGlow material which is applied to an inversesphere mesh to effectively simulate a simple point light (unshadowed diffuse normal lighting only). The material is translucent, unlit with “Disable Depth Check”. The inverse sphere mesh basically is just a way to mask which pixels should be affected and it’s origin is used as the origin of the “light” in the material calculations.
Despite 3 big explosions in the demo there seems to be only one use of this in the demo as released (around the fire-hydrant explosion near the start) . There are references to it being used in muzzle flash effects but that seems to have been turned off. Finally there seems to be a movable point light at the same location.
So why was this used and only used once? Is this technique significantly better performing than a non-shadowing movable point light? Why was it disabled elsewhere as released?
Having made something similar, it was much more expensive than simply using a light with shadows disabled and min roughness set to 1. The only benefit it grants that I can think of is lighting an entire model rather than just the normals facing the pivot. If you’re ever only going to have 1 in a scene, then the performance impact is minimal but more than 1 and you’d be better off using lights.
Thanks for your response. I would have been surprised it were faster than a light.
By “lighting an entire model rather than just the normals facing the pivot” do you mean something like adding an ambient value to those pixels of the model with normals not facing the pivot?
