For some reason the Virtual Shadow behaves like this on Transculent materials..
What is this ? Is it an engine bug ? Or wrong settings ?
UE 5.5.3
Lumen
Virtual Shadow Maps
Simple Optimized Water Material ( Forward Shading)
Video:
Hello there @Parazso!
That’s a peculiar behaviour you are encountering, it seems to be triggered by either the viewing distance or the angle in which shadows are observed.
As a first step, please check that ray traced shadows are disable, since you are working with lumen and VT:
If the issue persists after that, I would implement the steps found in the following guide, these were quiet helpful in a similar past problem:
Hi @Parazso,
Hmm, this is very odd to see because the pattern of when pixelation occurs isn’t in-line with how Virtual Shadow maps are usually rendered. Normally shadow LOD level is determined by distance from the camera, not the frustum, so you shouldn’t be seeing a difference in quality when changing your camera angle.
The pixelation pattern makes me wonder if it’s something with your page caching. Try using the command r.Shadow.Virtual.Cache 0 just for debugging. This will kill performance but force the pages to redraw every frame, eliminating this as a potential cause. (Given that VSM works best with Nanite and I’m guessing your water is translucent non-Nanite, this would check out).
If you’re still seeing this issue, I would recommend using the Virtual Shadow debugging visualizers to look for any funny business.
Send me anything that looks off, and we can go from there to narrow down the cause and solution.
Also, here is the official VSM documentation for reference.
Let me know how that goes!
In your project folder open config/DefaultEngine.ini
Under the [/Script/Engine.RendererSettings] section, add:
r.Water.SingleLayer.DepthPrepass=1
r.Water.SingleLayer.ShadersSupportVSMFiltering=1
r.Water.SingleLayer.VSMFiltering=1
Restart your project.
Thank you all for the tips, I finally found the problem.It was not what I thought, the whole material was malfunctioning..
I transferred it to Transculencyvolume, this solved the problem, but the reflection of the directional light from the water disappeared
Thank you, this is a plain optimized material plane. I assume your solution is for the use of normal water ?