Miarmy Crowd Simulation Documentation (English)
Level of Details Setup
General
Level of Detail (LOD) is an optimization for rendering and it can make our render much faster in many cases.
LOD responsible for separating the near and far space in camera frustum, and it will tell the renderer choose the high resolution geometries if the agents are near the camera, whereas choose the low resolution geometries if the agents are far enough from the camera.
Instructions
It currently support:
Render Method | Support LOD |
---|---|
RenderMan | Yes |
Arnold | Yes |
V-Ray | Yes |
Redshift | Yes |
Mesh Drive | NO |
Level of detail
Due to all of our render system support Procedural Primitive (DSO) now, we can let the render engine choose the right geometry in render time.
Below image shows that frustum has been split into 2 spaces, near(area 1) and far(area 2). If the agents are in area 1. Our renderer will choose the high resolution geometries, otherwise it will choose the low resolution geometries.
Area 1 choose hi-res geometries, Area 2 choose lo-res geometries
LOD General
For setup the LOD, there are 3 steps:
1. Put the low/high resolution geometries at right place
If you already read the previous session, Skinning and the Simple Geometry Randomization, you may notice that our geometries should be put into the "Geometry_<Agent Type>" Group and arranged. But our low resolution should be put into the group which name is "LODFar_casual"
The group of hi-res geometries and lo-res geometries
But if there is no such group (because you're using the old file from previously version), please click below button and add that group
2. Skin the geometries to Original Agent
Also, the low resolution geometries have to be skinned onto the original agent.
3. Setup camera distance and make them separated parts.
Select the LODFar_XXX node and directly set the value "moreThanDistance" to desired value
more than 250 from render cam, it will use lo-res geometries
LOD Render Test
Basefount Technology