Some progress: the grid, selection, wireframe, light gizmos, reverse Z-buffer, path-tracer/raster compositing and a few more bits. It's all coming back together.

Do you have any advice on how to handle occlusion when rendering a mix of gizmos and ray traced content? I’ve hit that issue recently with Imguizmo, and am itching to upgrade… but I’m not sure how gizmos should look when viewed through glass, behind objects, etc

I render manipulators always on top since it’s high chances the manipulator is inside the selected object. Other stuff, like spotlight cones, I pass the pathtraced depth buffer to the raster pipeline as a texture, then convert the fragment.z to depth and discard (or fade). That won’t handle refractions or fractional opacity of course.

Going multichain: which is the best chain to start?

Looks great

Can I ask if you will open-source this project?

With my current workload, not very likely.

Wish light manipulators looked that good in Solaris or Katana. Seeing the shape of a rectangular light spread is so useful yet never implemented in DCCs. And you even have round corners! Awesomeness

Cheers, it’s round corners because the rect light spot is a convolution of a cone over a rectangle, so I went for the exact shape instead of an approximation. Some use a super ellipses shape, which I don’t like when scaled: round corners are more elegant to me.