Hi, I've recently returned as a corona user but am having a few issues. Can anyone help?
1. When using the shadow catcher material with alpha mode set to compositing, i'm getting very stepped gradients in the alpha when saving as png (rgb 48 bit) exr 16 bit half float seems ok ...think windows 256 colours. Direct visibility override is set to black.
It's also incredibly slow to render, i'm talking at least 50 times slower to reach a similar noise level than if I render the floor with a standard corona material instead of the shadow catcher. I've spoken with Marcin today who recommended increasing gi vs aa and light samples, but that didn't really make much difference apart from better balancing the direct/indirect noise levels. The scene is lit by a single hdri which i've tried clamping in the output but that made no difference. All of the albedo's are fine too. Most of the noise seems to be in the shadows (judging by the cs shadows pass). The scene is very simple, consisting of 5 materials:
shiny silicone with, diffuse set to 0, reflectivity set to 1, ior set to 1.6, glossiness set to 1, translucensy set to 1 transluceny colour white (value 188), absorption white (value 188) distance 1mm.
aluminium, diffuse 0, reflectivity 1, ior 999 (as per new pbr instructions), glossiness 0.4
plastic reflectivity 1,ior 1.6, glosiness .8.
translucent plastic, virtually the same as the silicone apart from absorption colour/distance.
shadow catcher, alpha for compositing, enviro backplate set to black with no map.
hdri image used as an envornment map for lighting/reflections.
I'd share the scene, but i'm under NDA.
Question 2: Is there an exhaustive help file? I can find tutorials, but I could do with a proper reference of all features/dialogs/ render options/materials/lights etc so I can get back up to speed more quickly when the tooltips are insufficient.
Final question (unless question 2 is a yes): in the stats tab of the vfb at the bottom there are sections for GI caches and performance. What does this data actually mean, and how can I use it to analyse and correctly configure the renderer on a per image basis?
Cheers
Michael