Original: an old photo I took in the forest
3D voxelized scene in the day.
3D voxelized scene in the night.
3D voxelized scene in the day with volumetric fog.
This was a fun quick experiment:
Using an old photo I took in the forest, I used ml-sharp to generate gaussian splats from it, voxelized it in my toy renderer, relit the scene dramatically.
#rendeting #metalrt
Path-traced render of an off-white folded fractal.
Path-traced render of a colorful folded fractal.
Coloring a fractal is interesting in itself. Finding good colors is tricky. Applying them to the "right parts" is tricky.
Here's a comparison of off-white vs a value that's displaced relative to the folding for each iteration (then mapped to a color ramp).
#fractals #rendering #metalrt
Render a fractal in 2 easy steps.
1. Render a sphere.
2. Render the rest of the... fractal.
#rendering #metalrt
How to render a fractal in two easy steps:
1. Render a fractal.
2. Fold it into a fractal.
#rendering #metalrt
Fractal mitosis 4.
#metalrt #rendering
Fractal mitosis 3.
#metalrt #rendering
Fractal mitosis 2.
#metalrt #rendering
Fractal mitosis.
#metalrt #rendering
Cem Yuksel's update to the teapot model looks great with a glass material!
#metalrt #rendering
Naive attempt at loading the Zorah scene. Uses 200GB of RAM. With enough swap and patience, who needs optimization?..
#metalrt #rendering
Morphing fractal shapes.
This is a Julia fractal with a linear interpolation of the distance function between iteration counts. Starts by interpolating between 1 and 3 iterations, then 3 and 5, etc...
The material also changes for no particular reason...
#metalrt #rendering #fractal
Fractal SDF rendered as voxel spheres.
The high resolution one has close to 1 billion spheres.
#metalrt #rendering
Fractal parked in front of the Bistro, day time. Path traced in my toy renderer.
Fractal parked in front of the Bistro, at night. Path traced in my toy renderer.
Fractals, in situ.
#metalrt #rendering #FractalFriday
Fractal with a translucent pink material, planted in the Sponza scene. Path-traced in my toy renderer.
Fractal with a pearlescent material, planted in the Sponza scene. Path-traced in my toy renderer.
Fractal with a foamy material, planted in the Sponza scene. Path-traced in my toy renderer.
Fractal under a spotlight, planted in the Sponza scene. Path-traced in my toy renderer.
Fractals, in situ.
#metalrt #rendering #FractalFriday
@jbakergraphics.bsky.social has an amazing collection of fractal distance estimators! jbaker.graphics/writings/DEC...
All the authors are listed on the page. There are some gems in there!
#metalrt #rendering
Quick test of participating medium - simplest possible: on primary rays, N% chance of bouncing in a random direction at an arbitrary point between the origin and the intersection.
Not a bad result, but I'm disappointed not to see better refracted light trails under the sphere.
#metalrt #rendering
Nice little palate cleanser for my mind: easy renders of a few landmarks using OSM data.
Some of these are 100M triangles. Plenty of detail in there.
#metalrt #rendering
A single small area light on the back wall is shooting photons into the scene, bouncing through the colored glass dragon and hitting the walls and floor. This is showing a trace of the path of each photon.
A single small area light on the back wall is shooting photons into the scene, bouncing through the colored glass dragon and hitting the walls and floor. This is showing just the final photon positions and flux.
Sponza scene, illuminated by 'physical' photons shot from a point light inside a colored glass dragon.
Sponza scene, illuminated by 'physical' photons shot from a point light inside a colored glass dragon. Showing the traced paths.
The photon debugging view is working pretty well now. I can navigate the scene to look at the paths, bounces and final photon positions - along with a preview of their impact on the scene.
#metalrt #rendering
Here's an idea: shoot some photons and add them to the scene as actual emissive spheres to illuminate it.
Couple of drawbacks: it's very slow and it's not very effective.
#metalrt #rendering
Photons are shot from two point lights (a blue and a red one). They are recorded in a buffer, then added to the scene BVH as spheres and thin cylinders, with emission from the flux.
Photons are shot from the area light. The distribution is horrendous and needs a little work.
Photons are shot from the area light. The distribution is tweaked a bit but still bad.
Want to do a quick photon mapping test, but it's a slow start because I spent most of the time so far playing with the debug rays instead. It does immediately show that the basic GPU hash rand I used is no good.
#metalrt #rendering
1/15) 🧵 Breaking down a quick experiment: Simple but slow displacement for ray-traced triangles
#MetalRT #rendering
Top-left: Typical path-tracer with a simple triangle-based model.
Other three: Same model input, with increasing displacement added while tracing the rays.
Subsurface scattering render of a cartoonish bust that looks nothing like me.
Diffuse render.
Marble render.
Subsurface scattering render of the bust of David. This one came out ok.
"Self portrait" that looks nothing like me, as a test of a simple subsurface scattering implementation. I was aiming for the classic marble look, but there's plenty of room for improvement...
#metalrt #rendering
Halvorsen attractor.
#metalrt #rendering
Glass Lorenz attractor.
#metalrt #rendering
Simple variant of the Lorenz strange attractor.
#MetalRT #rendering
