That looks really bad with the contrast

[frosty sparrow]

What does it look like now, where are the lights and what are their properties

[stoic agate]

rather than raycasting and calculating where exactly I should place my point light on the beam

I just literally placed some point lights along the beam and moved along the terrain

to better see where they should bounce, right?

I can show you with and without area lights

without point lights

with point lights

1 more example

without

with

it looks better but you can see the floating point lights

also it's still weird

I might need colorbleed too

maybe I should use one point light with a very large diameter and somehow decide to place it one of the raycast hit points

[frosty sparrow]

Your point lights have a very small radius so that prevents them from looking like ambient light

[stoic agate]

they have a range of 50m 😄

[frosty sparrow]

radius, not range

[stoic agate]

oooooh

I'm stupid

oh my

[stoic agate]

goddamn

I missed this

IQ in the single digits

[frosty sparrow]

In my experimenting this one point light looks great anywhere relatively close to the surface that's being hit, as long as it doesn't go through it

Might still need a couple of more of them for situations where the light in motion hits different objects at different parts of the cone

[stoic agate]

yes exactly

[frosty sparrow]

the tricky part is probably to do this but make it smooth while the lamp is being waved around

[stoic agate]

but it seems like intenstiy needs to be high

for the light

I mean the point lights

[frosty sparrow]

Their intensity should be proportional to how much light is hitting the surface

[stoic agate]

yours looks good

what's your radius setting?

[frosty sparrow]

2.5 in the further away one, 0.9 in the closer up one

Intensities of 34 and 47 respectively

[stoic agate]

in lumen??

[frosty sparrow]

Yes

[stoic agate]

better

but my intensities are way higher

[frosty sparrow]

Intensities are relative to exposure in HDRP

[stoic agate]

also the tricky thing is I can focus the beam of mine with the mouse scroll

[frosty sparrow]

The smaller the area of the light is on the surface that's being hit, the smaller the radius of the "bounce" light should be as well

My first option would be to do say five raycasts for five lights around the cone, halfway from its center, accounting for change in the cone

Then at each raycast place the bounce light source at maybe 25% of the distance in

Then see how it looks and figure out how the radius and intensities need to change over distance

[stoic agate]

I hate the fact that one of the raycasts might not hit

maybe I could find overlaps or something

like if I place a trigger on my spot light cone

maybe too expensive every frame

[frosty sparrow]

Eh
Try a spherecast instead if you need a "thicker" raycast

[stoic agate]

but I guess I'd be able to calculate the surface or something

[frosty sparrow]

But missing one would be intentional

[stoic agate]

I'm getting ahead of myself now

[frosty sparrow]

If a raycast misses it presumes that that part of the light cone is missing the thing you're illuminating, so less light is bounced

Which is exactly what we want

[stoic agate]

I see what you mean

[frosty sparrow]

Even with just one raycast down the middle it might be plenty convincing enough, who knows

[stoic agate]

more raycast hits = higher surface = more point lights

but like, If I'm just hitting a wall in the upper right corner with my beam, it wouldn't hit if the wall doesn't rotate in

but that's a specific case now

shouldn't get ahead of myself

also maybe a colorbleed effect or something

[frosty sparrow]

You can do as many raycasts and lights as your performance budget allows, but only so many

[stoic agate]

ok ok

thanks mr. pro

I will experiment further

I'm still in my prototype phase, just testing

but I need to figure these things out

btw should the point lights not use shadows?

I don't know how they'd mix with the spot light

I tried and it looks alright either way...idk

[frosty sparrow]

They're significantly more expensive with shadows, and with a large radius the shadows are hard to notice
But they technically do prevent illuminating areas that should not be illuminated

[stoic agate]

yes but it's FPS so I don't know if I could really tell the difference

[frosty sparrow]

The issue with color bleed is that raycasts are against colliders, not meshes with textures on them
Iirc you can also cast against meshes, or get the nearest mesh vertex and UV coordinate, but the first option is way more expensive and the second one is way more complicated
You could instead just get the meshrenderer of the collider hit and read a color property stored in its material or somewhere, to guess the bounced light color

[stoic agate]

they're going to be outside my vision, those areas, no?

[frosty sparrow]

Usually

I wouldn't use the shadows unless there was some absolute need for it

[stoic agate]

ok I won't

I almost forgot, I have lit particles flying around

it might look good with point lights

[frosty sparrow]

Yes, but since they can get very close to the lights you must ensure the radius is never too low

Or the particles get logarithmically brighter when approaching the light

[stoic agate]

radius of what? the particle system

yeah, because I want them to move in world space, the radius is giant lol

[frosty sparrow]

Radius of the bounce lights
The smaller the radius, the more the light's brightness is concentrated at its origin point

You can test it by setting a light's radius to 0 and moving something very close to it
I expect it will practically nuke the screen brightness

[frosty sparrow]

This also often requires optimization
If you've got some kind of atmospheric specks, like marine snow in your case, you don't want to cover the whole level with it, but rather have the system follow the player to only spawn the particles in their vicinity

[stoic agate]

silt but yeah I wanted to avoid the particles moving with the player

I see that all the time in horror games with dust flying around

I'll see how I can get it to perform

[frosty sparrow]

They don't move with the player if if the simulation space is world, regardless of how the system moves

[stoic agate]

hmmm it might work if I have a big enough radius

I didn't want them to just spawn into existene

and I didn't want them to spawn from the player direction but I guess I make make the shape a sphere and spawn them randomly from within

that might work

[frosty sparrow]

Fading them in/out is pretty much guaranteed to be unnoticeable in that situation

The advantage of following the player is that you don't have to deal with the culling to keep the particle counts sane

[stoic agate]

that's true

ok more thing and then I'll leave you alone

I'm thinking about how exactly I'll make my cave system

I am using the digger asset but its shader is broken in unity 6, so the test cave mesh you saw was an exported fbx mesh

I was going to texture exported meshes with polybrush or something

is this an ok workflow?

though I think the cave meshes may have broken lighting because they lack UVs

I'm just considering the way I should approach this

I briefly considered importing all the meshes in blender, generating UVs, importing into substance painter and texturing them but my game would be like 200GB

I know splat maps are the way to go, maybe I hsould just go with exported digger meshes + polybruh and see if the polybrush shader fixes the shadow issues

[frosty sparrow]

Nnot really
Lightmap UVs are required for baked lighting which can be generated automatically
And I wouldn't ever consider unwrapping such an organic mesh with polybrush, or in general

You'll probably want a triplanar shader

[stoic agate]

I thought polybrush can do triplanar

someone told me that microsplat shader still works for digger in unity 6 so I guess I'll bite the bullet and buy all those needed add ons for that if it will make my life easier

I think polybrush can only splat 4 materials anyway which is a few too few for me

and I think I can make the bounce light look good with some raycasts, then sample pixel color and turn the light color to that of the pixel

I will experiment

[frosty sparrow]

Polybrush is a mesh editing tool that can edit UVs
Triplanar is a shader that doesn't require UVs, it doesn't matter where it comes from as long as it supports the shader features you need

You may find that "sampling a pixel color" can be quite a complex that
I would recommend storing a color in some component or asset instead, if possible

[stoic agate]

Now I'm curious lol

I will see how that works out first

I didn't know polybrush edited UVs at all