The future of real-time NPR shading relies on the ability to render in 3D, apply 2D effects, use those effects to alter the 3D render, apply 2D effects, ad infinitum... The industry is stuck in the "Render, composite, done." mindset. #b3d #bnpr
1*
1*
Opening up this process would open up 3D looks that would revolutionize Animation and Art in general. It is something I hope for in #blender #Eevee. And I hope the developers lean in to this possibility in the future.
2*
2*
This would also take advantage of Eevee's speed and unique abilities (3D->2D->3D) as complementary to #Cyclesx (its speed and denoising is making Eevee less and less relevant for PBR workflows).
3*
3*
I'll give a concrete example - something simple, you want a white square to represent the highlight of a sphere. Right now, this is an almost impossible task. But not impossible to rationalize...
4*
4*
a.) Find brightest spot on shaded sphere
b.) Calculate UV location in screen space
c.) Map white square to that location
d.) Composite with sphere shader that has no highlight
5*
b.) Calculate UV location in screen space
c.) Map white square to that location
d.) Composite with sphere shader that has no highlight
5*
These calculations are much less intensive than some of the other effects already in Eevee. It all has to do with breaking the "Render, composite, done" mindset. And I think a "Render, composite, render..." workflow is much more freeing and revolutionary.
6*
6*
Kudos to the developers in the 3D industry. These are just my thoughts after trying to make (and break :-)) things for years. We haven't even scratched the surface yet.
PS - Softwares, please meet with @RafAnzovin and give him everything he wants :-D
THANKS! END
PS - Softwares, please meet with @RafAnzovin and give him everything he wants :-D
THANKS! END
• • •
Missing some Tweet in this thread? You can try to
force a refresh
