Hi there! I'm trying to implement chapter 4 of the Lighting and Shading section in GPU Pro 5. Basically, how to optimize my screen space reflections using a mip-mapped Z buffer to quickly converge on the intersection point of my reflection ray.
Sadly, the author wasn't allowed to release the code/demo he talks about in the article, so I've had to work out most of the shader myself. I'm close, but one thing I don't get - if you are always starting at a lower mip in the HiZ buffer, won't your starting ray depth often (always?) be _behind_ (greater Z) than what you read from HiZ buffer? because you make the HiZ buffer taking the min(...) of the more detailed mip.
If you've implemented that chapter, or even just read and understood it, or using HiZ tracing before, I'd be curious to hear your thoughts.