kobra VS pbrt-v3

I have been working on a rendering engine for quite some months now. It primarily focuses on raytracing with OptiX (later plan to add GLSL/compute shader and Vulkan RTX) and is oriented towards advanced research in graphics. I would like to hear what others think of this project, and what they would like to see from something like this.

I've been working on a path tracer for the past few weeks, and I think I've got it to a presentable state. However, I feel like there is still a lot I could do, both in terms of image quality and performance. Any tips or cool things I could implement?

Physically Based Rendering[0] was an excellent textbook when I read it ages ago and conveniently enough it looks to have been updated with a new edition last year.

[0]: https://pbrt.org/

Physically Based Raytracing

Both of these books are free, and both are written in C++, but they can be done in any language. The first book, a raytracer in a weekend, is part of a series, you can find it here: https://raytracing.github.io/ And, if you get to the third book in that series, or you need a reference book, the PBRT book covers the math in more depth and discusses the latest theory, you can get the last edition of the book (5 years out of date) for free though: https://pbrt.org/

Making a “complete” one is a never-ending rabbit hole you can spend a lifetime on and is a very active area of research covering more advanced geometry, probability, optics, machine learning etc etc. A great introduction to that is https://pbrt.org

This isn't the highest priority but if you haven't already, learn how computer graphics works. Get a working knowledge of the camera matrix, real time graphics (say, OpenGL but threeJS is an option), and photorealistic graphics. PBRT is the go-to for photorealistic graphics. The first two books of Foundations of Game Engine Development are way more useful than they have any right to be (and my favorite textbooks I've ever read, 10/10).

Org-mode strength though is in working with different languages in a same source file, which I am not sure if Knuths version does. Anyway, to see how the original idea looks like, check the Wikipedia article, or to see it in real-life see some of books that are written in the literate style, like Physically Based Rendering, which seems to be available for free nowadays or C Interfaces and Implementations.

I contributed a tiny bit to pbrt[1], and one of the things I loved was that if you just plugged in physical values you almost always got great results with minimal tweaking.

The Octane data seems most complete at first glance (with complex IOR etc), but for things like milk and blood I expected at the very least some absorption coefficient for the translucency or similar.

[1]: https://pbrt.org/