Since there is no detailed documentation out yet, not that I could find anyway, I wanted to see if anyone knew more about the options in the lightmapping settings.
Precompute
I understand what this does but I have to guess what the Resolution and System Size values do. Some more detailed knowledge would be nice.
Baking
The modes, what are they? Are they sorted by prettiest to fastest or are they to be used in different situations?
Resolution here is a bit more straight forward and I’m guessing padding has something to do with the seams? Atlas size is pretty obvious though and so is the sliders Direct, Indirect and AO exponent.
Runtime
Albedo boost and Indirect makes sense and its easy to see what they do.
CPU usage is probably settings for CPU usage. Yeah…
Environment
Also pretty easy to figure out imo.
There might be stuff I find obvious that someone has no clue about and the stuff I’m unsure about might be basic level knowledge for some. Lets help eachother out ^^
Precompute
Realtime GI is applied to static geometry via lightmaps. The resolution property is the global scale that controls how big texels are. The value of 1 means that a texel will be 1 world unit big (so as big as the face of the default unit cube).
System is a grouping of objects. System size defines how big is the texture (the width and height) into which objects get packed. If you lower this value more systems will be created. If you increase it fewer ones will be created.
Resolution can also be increased or decreased per system by using lightmap parameters (see the Object tab).
You can see how many different systems you have and which object belongs to which one in the “GI - Systems” scene view mode.
Generally having more smaller systems is encouraged as then a certain system has higher chances of depending on smaller amount of nearby geometry. Fewer dependencies mean shorter precompute time.
Baking
The modes available are now under development, but generally there’s the non-directional one which causes surfaces receiving lightmaps to just be diffuse and loose normal mapping. Then there’s the directional one (or ones, we’ll see what we will ship with) that stores some information on the direction from which the light came and how directional (or ambient) it was and also stores the intensity differently. That allows us to evaluate a more proper BRDF and get bump mapping, specular, etc.
BTW those properties are still getting a small makeover, so expect changes in the upcoming betas. Docs will be added once we are closer to the final look.
