Sample Count Image for Nov. 10th Submission.

Rendering Light Source and Added Multiple Scattering. Refractions are still not correct.

Getting backface hits around edges [magenta is a backface hit]. Refractive material is not fresnel-reflecting with microfacet normals. Reflections are incorrect at the point where 2 spheres touch.

With the teapot. Adaptive sampling with step function from 8spp - 32spp. Using QMC sampling for perspective-d lambertian spherical light. Time: 4min, 54sec. Next, glossy reflections.

Fixed the layers. Lovely soft shadows [without the teapot] using true Monte-Carlo sampling. Adaptive sampling - step function from 8spp to 64spp.

Fixed direction issue. Shadow is sill layered. On to making the light perspective instead of orthographic.

With the teapot. It is backface culled, but will fix that later.

First attempt, wrong light direction. Shadow is layered. No glossy yet. Orthographic light not computing actual disk radius. True Monte-Carlo Sampling. Fixed # of shadow samples to 16 for simplicity. Forgot the teapot!!!

No soft shadows. Image for comparison.