&learningjam2024
Back on the PBR demo. Went off a tangent and implemented my own local "Shadertoy", so that I can use my preferred text editor without having to do additional copy-pasting. I've written a blog post about it.
https://unlitart.com/Blog/#PBR%20Demo%20WIP%20and%20Writing%20Offline%20Shadertoy

&learningjam2024
As an optimization I incorporated some ray tracing into the ray marching routine.
First we trace a bounding box covering the objects and also trace the ground plane. If we miss the box then there's no need to ray march anything and the results of the plane trace is returned.
In the case of a box hit, the ray march will commence from the entry point until minimum distance or the exit point has been reached. In the case of the latter, the plane trace is returned as well.

Unsure about the performance gain. In standard Shadertoy editor mode it seems to be over 10% (according to Psensor readout) but less than that in full screen. At the very least I got a clean trace of the ground plane now.

&learningjam2024
So I managed to implement raytracing (in Shadertoy). Gets expensive really fast but I'm quite pleased by the results. Rough materials are still too noisy, and I think the low sample count plays the major part unfortunately.
Sampling 4 times per sub pixel, sample count is halved for each bounce, 2x2 AA and 2 bounces. Last bounce always samples the environment light.

&learningjam2024
Seems like I've addressed the bugs I had. Also found an updated version of the paper about sampling the microfacet normal.
https://jcgt.org/published/0007/04/01/paper.pdf
Currently doing 8 samples per sub-pixel with 2x2 MSAA which effectively makes it 32 samples per pixel. The variance on rough materials are too high for my liking but seems like it could be improved with a more sensible lighting scheme.

&learningjam2024
Making progress! All the pieces in place and more or less working.
Next I'll be investigating the black spots that are most apparent in low roughness materials. Also reflections are scaled up a lot on low roughness flat surfaces (like the floor). I suspect that my sampling is biased in some way.

(The funky environment light is for testing purposes)

Found a very good introductory blog post about importance sampling that cleared up some confusion I had.
https://patapom.com/blog/Math/ImportanceSampling/

Also found a newer paper going into further detail about the sampling scheme I'm using. Haven't read it yet.
https://jcgt.org/published/0007/04/01/paper.pdf

Currently sampling 32 times per pixel.

&learningjam2024
Here's the demo in its current state, still using direct lighting.

&learningjam2024
PBR demo is taking its sweet time cause I decided that I wanted to reflect the environment (just some procedural sky sphere for now) thus needed to learn about importance sampling, so most of the time dedicated to this project has been spent reading a bunch, again. It's been the most difficult subject so far for various reasons but things are starting to converge in my mind.
Tonight I wrote a program to test a function for generating micro-normals as presented in this paper: https://hal.science/hal-01509746/document
Seems to be working. Next is to get the other parts of the routine working.

!til &learningjam2024
Just finished processing this paper. I found it very informative but also hard to follow since the math notation made the equations look the same. My notes are basically a condensed version of the whole text, including all the 131 equations of which I'll review and color code the variables.
https://jcgt.org/published/0003/02/03/paper.pdf

&learningjam2024
Recently finished reading through the paper [PM] I mentioned in my last post. It certainly clarified some things and my summary has some inaccuracies in it.
Until I get around to address that I'd like to mention that equation (3), which I got from the Disney paper [DS], also has to be multiplied with a "normalization" term which is shown at the top of page 25 (eq. 3, the left fraction on the right side). This wasn't made clear to me until it was mentioned in [PM].
This is my usual experience with reading papers of this kind. Often I need multiple sources in order to fill in the gaps of my understanding.

Testing in my Shadertoy demo, the specular highlight (D-term) now behaves more or less as I would expect. Still got aliasing at the penumbra which appears to be due to the G-term (which handles micro facet shadowing), so that's what I'll be working on next.

!til &reveng-ann As a sign-off, for the teaching portion decided to majorly revise the first part of my writings (on solving Caesar ciphers using a transformer) to ensure overall quality. Gave a project recap and had a fun time!