Regarding Exponential Height Fog itself regardless of Volumetric Fog being enabled:
One thing I don’t understand is that why Exponential Height Fog is already lit by itself while there is no Directional or Skylight placed in the level?
A realistic approach would be for the fog to be lit by Directional and Skylight. It’s brightness should be bound to Directional and Skylight intensity, and it’s color should be bound to skylight color. That’s a very simple, yet very nice step towards having a more physically based fog. Artists will be able to achieve more realistic results. (While having overrides for brightness and color for artists to be able to achieve stylized settings as well).
Exp fog isn’t even PBR, so I’m not sure that would do much. It would just be skylight * directional * fogcolor * densityatpixel, instead of fogcolor * densityatpixel. You could do the multiplication yourself. Edit: I’m easily distracted by my own misconceptions. What I mean to say is, it would be great to have some PBR “love” given to exp fog. I’m not sure if basic light interaction in particular would be especially useful, since exp fog doesn’t benefit from occlusion like vol fog does, but there are likely other ways to take advantage of it without going all the way towards vol fog’s expensive implementation, and it would be fun to brainstorm such ideas.
I suppose I should ask, how just how closely knit is vol fog to exp fog? They seem like completely independent features that are just tied to the same actor for convenience.
The way Volumetric Height Fog works is that you don’t see the fog in shadowed parts. However, when placing vol fog as a particle sphere, you still get to see the fog in the shadowed parts too and that’s not how it works in real life. Am I missing something?
Notice how the pools of direct light are blown out. The light source is bright enough that even a tiny amount of fog (like you’d expect to see in a cave) can create visible shafts.
Here’s something made in unreal:
The Volumetric scattering intensity is at its default value of 1.0, and the exponential height fog density is 0.5
If you want even thinner fog, you can use particles to model where airborne particles should appear. If you create a particle system which spawns at the ceiling crack and drifts down (simulating sand/dust/pollen from outside) you can eliminate the fog entirely.
Exposure and light intensity should have the same effect. If you light value can’t change for whatever reason, just increase your EV and try to match reference.
Using the exponential height fog it works fine. But my issue is with particles. I’ve already placed a particle entity inside the cave and the problem is that I continue to see fog in the shadowed parts.
Expected behavior would be to see the fog only in the lit area (below the crack only). However, as you can see hereI’m seeing dense fog everywhere in the shadows as well. Fog shadows seem to be not working properly with particles.
Does anyone know if there’s a way to maintain fog density on small particles? I was setting up a rolling fog particle similar to 's one from his Shaderbits blog (link). He mentions that a small, camera-linked sprite would be better for performance than a gigantic one encompassing the entire level. However, from all of my experiments, it seems that the density of the fog is directly affected by the size of the particle, even when it’s not sampling the particle radius. Here’s the bit I’m specifically trying to deal with from his blog post:
But shrinking the particle size down to something that would fit just in front of the camera (or even roughly human scale at around 200uu) caused the system to fade out and lose all the interesting density. To make sure I wasn’t doing something screwy, I ran a test with an extremely simple fog particle material with an albedo of 1 and extinction of 0.5, attached it to my camera, and then simply scaled the particle up and down. What I confirmed is that even if the particle covers the entire camera frustum, the apparent density of the effect as it recedes into the distance is directly linked to the scale of the particle (set in either Cascade through the size node or in the editor through simple scaling of the particle system - scaling either responds identically). See the shot of the material below - I’m specifically not using the ParticleRadius node to influence the density, theoretically this material should be showing the same fog density so long as the camera is looking through it.
This basically prohibits the notion of creating a tiny camera-linked particle system that fakes the effect of a much larger (and less performant) one. I’m not sure if this is a bug, a limitation with the current implementation of Unreal’s volumetrics tech, or something that can be remedied in some way. Here are the shots detailing my experiments trying to isolate the issue:
And here’s the particle system with the particle size of 1024 for it’s default 1.0 scale (ideally this would only be anywhere from 64 to 256uu in size if sitting just in front of the camera):
EDIT: Unrelated to above, but I’m super happy to hear the moveable light reflection capture issue sounds like it’s being addressed!
+1 Same deal with the project I’m working on.
Thank goodness! This will be a lot cleaner than having to constantly drop in cube capture actors, capture, save out those captures, then manually plug them into reflection actors, and then do it all over again every time we update anything just because our lights are moveable.
This image Volumtric Scattering Intensity=2,The volumteric fog can make the rock brightness,But light no direct irradiate the rock
This image Volumtric Scattering Intensity=0 So the light back rock is black
So i want to know this is bug? or some way to solve this problem
Looks like static shadowing due to a Stationary Directional Light. Static shadowing doesn’t have a ton of precision and leaks like that. You can use Cascaded Shadow Maps near the camera to get more precision. (set DynamicShadowDistanceStationaryLight)
I haven’t yet seen anyone answer as to whether this is possible or not. I myself am working on a project where it would be very nice to be able to contain a fog effect within an arbitrary mesh, as opposed to just a sphere.
A good example of this effect would be a supported way to create the interesting car-shaped fog in the feature highlight video (which from what I understand–which is very little mind you, I’m new to this–is a bug)
