Graphics Programming weekly - Issue 68 — January 13, 2019

• start of series for programmers with D3D11 experience starting to learn D3D12
• explains what command queues are, how they express different GPU execution timelines and how fences allow synchronization between CPU and GPU

• explains how to render water using WebGL, Rust, and WebAssembly
• water supports reflection and refraction
• reflections are rendered using a reflection plane technique, that mirrors the world and clips fragments if they are below the water plane

• shows how to use screen space derivatives to create hard edges on a mesh in the pixel shader
• how to calculate greyscale from colors

• video explaining the rendering equation in an easy to understand and visual way

• list of common patterns for shadertoy programming
• included are: normalizing screen coordinates, color management (hue, sRGB), line drawing, antialiasing, blending, vector math, and random value generation

• preprint of the article to appear in GPU Zen 2 talking about the implementation details (C++ and GLSL)
• replaces the 3D histogram transformation with three 1D histogram transformations
• improves quality of the result under mipmapping and shows how to generate correct results with compressed texture formats

• walkthrough of steps to learn graphics programming
• recommends starting that beginners start with a raytracer/rasterizer from scratch before starting to use any graphics API
• list of project ideas for beginners

• explains how to rasterize the glyphs from a font into a texture. Providing the coverage information that is required by the technique discussed in part 1 and 2
• includes code for the OpenGL compute shader
• overview of typography terms and explain how to correctly position the glyphs to form the final text on the screen

• presents what vertex attributes are and how they are interpolated to deal with perspective distortions correctly
• uses this knowledge to interpolate normals and UV coordinates for the sponza model

Part 1 of the article:

• explains how classical forward/deferred pipelines work and how tiled and clustered algorithms can be used to extend these ideas
• includes animations to show the concepts in a more visual way
• presents strengths and weaknesses for each technique

Part 2 of the article:

• presents the clustered forward shading implementation
• how to subdivide the camera frustum into clusters and implement light culling
• provides links to further information sources

• frame breakdown of the Rise of the Tomb Raider with D3D12 on PC
• uses a light pre-pass architecture
• lens flares don’t use a visibility query but instead use the depth buffer to calculate visibility
• talks about shadow mapping, ambient occlusion, depth-aware upscaling, reflections, hair rendering, volumetrics, and tone mapping

• explains how to apply an expanding color effect that converts the world from greyscale back to color as post-processing effect using Unity scriptable render pipeline
• presents how to reconstruct world space position from the depth buffer

• shows how to set up a Docker container image that is able to run shader compilers
• supported are: DXC, FXC, GLSLang, AMD Radeon Graphics Analyzer, DXIL signing, SMOL/V and Vulkan SDK