Writing my PBR-Renderer as I proceed through the book

[Erfan Ahmadi]

Hello, 

I've been a rendering engineer for 2 years mostly focusing on Real-Time Rendering, Rendering Engine Design, Modern Graphics APIs, and GPU computation.

I have no prior experience or knowledge in Physically-Based Rendering although I've read some radiometry and I'm familiar with the concepts and the math involved. 

I want to begin writing my PBR renderer as I proceed through the book, (which is able to render pbrt-v3 scenes)

But It's unclear for me at which point in my learning/coding process I'm going to be able to have the knowledge to render the beautiful images in the PBRT book.

I would appreciate some wisdom on the best way to learn PBR and enjoy implementing my own renderer on the side. 

Thanks,

Erfan

[Un Flic]

Hello,

I'm doing the same. I think that it is best to read the chapters in order and not to skip things you think you know, because they show you how the concepts are implemented in pbrt. 

However, what I think you can skip on first read:

1. The chapters market as advanced content.

2. At first, for purposes of mastering techniques and easy debugging,  it is best to render simple shapes like spheres and planes. On first read you can skip intersections, sampling and general handling of all the shapes you are not interested in, including meshes and bounding volume hierarchies, also texturing and fancy cameras. IMHO the most important parts to grasp on first reading are the radiometric concepts, sampling, reflectance distribution functions and Monte Carlo integration. 

Then you can go back and add features.

[Un Flic]

And of course how Monte Carlo integration is used in the context of light transport.

[Erfan Ahmadi]

Thanks you so much for the details, this is really helpful