#stupid unity, again
1 messages · Page 1 of 1 (latest)
Unfortunately your problem is shading of the material surface, not shadows from the shadow casting
If your flashlight is very close to your camera, you could get away with disabling shadows entirely
And still have the same problem
HDRP Lit materials don't allow incoming light to "wrap" around to the dark side of the object, as that doesn't really happen in reality
Because normal maps simulate surface bumpiness, the "dark side" can appear even on geometry that you're directly looking at
I see what you mean
so what you are saying is
Toon shaders allow this but don't work great with HDRP's PBR lighting
there is no real way to reduce them other than reducing normal intensity
HDRP's translucent type materials might allow it also
What material properties do you have now and what does it look like?
ok you were right, neither shadow dimmer nor tint did anything for the normals
and I don't know what normal bias does but it seems to do nothing
it applies to both my rock meshes and a custom shader I made for polybrush
they're different materials
but I'll show the rock mesh properties
it's mostly just default hdrp lit
for cave walls, there's nothing much to see
normal intensity is the only thing that helps so I should stick to that I guess
Ah, I assumed you tried translucency with this
meh
I'll just finish the bounce light thing
then reduce normal intensities
what can a man do
sorry for going in circles
I tend to overthink
my september resolution is to stop overthinking overthing
and spending days on 1 thing because I can't figure out what the optimal way to continue is
and just DO
With our powers combined we'll get so little done few will believe it!
Lol
What about you
You making a game?
I definitely could be!
But I spend more time on instructor work
And squashing problems here is as good a way to practice as it is to procrastinate
So you're even lazier than me
You help work on other people's games
Instead of working on your own lol
You have the knowledge, you can make a game
Nothing seems to be there stopping me, but I've been leaving projects halfway for a very long time
At some point you have to face the facts and focus on doing what works and gives some kind of real results
I'm trying new things too though, like getting involved with teams and projects that give me more opportunities to do what I'm good at
The quickest way past a brick wall is not always through head first
@oak bloom oh and btw I did find out how to tweak the light spread, as well as how to reduce shadow effectiveness
Per material anyway
Subsurface scattering can make the light "spread" to the darker side, though at the cost of becoming blurrier and noisier
Translucency can reduce shadow strength, though using this might make just about all of your caves lose shadow effectiveness
Worth noting that real limestone shows a bit of both of these optical properties though
But in your cave the effect might be undesirable
I see
I also fully released a game solo
With very underwhelming results
I understand the reality
Everyone told me my new game idea is great, but we both know it's about the execution, not idea
But I don't care, I will try
Working in a team like you is probably objectively the best way to try
I assume I'd need a custom shader
When you think about it
It's an fps game with the only light source coming from the player
Am I ever going to see any shadows anyway?
Except normal shadows
Cause geometrically, I shouldn't really see shadows. Only very slighty because of the off-center light source
HDRP lit gives you many different material types, including translucent and subsurface+translucent
Possibly not
I mean, I can try, whatever
I've been wasting so much time just thinking about how and all
I just need to go with something
I suspect a good portion of the problem comes from the tiling of the texture
But nothing I can really do about that other than make my own with substance designer, but that's a whole new can of worms to open
This particular problem probably not
Tbqh the lighting you have already is perfectly adequate
Finetune it later when the rest of your game is nearing completion
The fact that it exists, not
But the shadows look low res because of the tiling
Even though it's a large texture
When you tile it at .1 or .2, it's still low res
But ye ok I will keep moving forward
That seems to imply to me that tiling is not the issue but
If your game gets some momentum you might end up replacing the whole cave mesh anyway
I will
But I'm doing this part for the vertical slice, it needs to look good too
Digger asset dev fixed his shader and I will look into that going forward
But that has nothing to do with the textures
Point is at this stage of development it's not reasonable to try to do everything in final quality since you probably have other priorities too
Yes that's what I took away anyway
When I said I need to stop overthinking
And get stuck on the same thing
Reminds me of the last game I published
I had a massive "unsolvable problem"
Would never be able to solve it
Deal breaker
So I took a month's break or something
I needed it mentally
I came back and solved it within 15 mins
And also another big issue lol
The placement looks good
More important that the resulting lighting is visually appealing
of course
thanks
stupid question, but it needs to bounce off in the opposite direction of the hit, right?
like, 1 meter or less away?
Maybe more if using point lights
If it's too close it'll just illuminate the surface your flashlight is pointing at
Spot lights facing away from the surface might work better in some circumstances
You're right about that
I need more def
Although spot light isn't a bad idea at all either
But I think small poijt light pribably better
Maybe
I don't know
I'd start with 1-2 units off the surface along the ray
That probably works well enough
1-2 units? You mean meters?
And I guess you mean on the opposite vector by along the ray
Or do you actually mean back on the same ray
Same thing ^^
Not sure what the difference of "opposite vector along the ray" and "back on the same ray" would be!
I mean right angle
As a real bounce
Or along the same vector backward
I would think right angle as bounce, yea?
Although I guess I don't know much about physics
It'd have to be a reflection vector against surface normal rather than a right angle
Probably better to just move it along the vector
Simulating a reflection probably would reveal the illusion since the fake bounce light would have to move around a whole lot
Oh yeah, I didn't consider that
Good point
It would probably be ver jarring with the light flipping due to normal directions
But then, if we exclude that
Why do you think it would look worse with right angle, rather than reflecting back
Wouldn't it still be "closer to reality"
I mean to say reflected rays simply aren't right angles
The idea of faking it this way is to illuminate the area in front of the lit surface very generally
Certain "inaccuracies" help conceal the fakeness of it and thus would look better
Because true bounce lighting is ultimately unachievable
At least without Enlighten or RTX
If using a point light, rotating and maybe also moving it in reflection vector direction might give you more correct fake bounce lighting for that particular ray
But I would expect it's still too finicky due to the hit normal changing very rapidly
And it could reveal how few rays and lights we have
That's why I think big vague point lights could be better
This seems like an overthink though ^^
I understand what you mean
You might be right
But I'm afraid, for example, hitting a wall at a very small angle looking bad without reflective vectors
If shine close to 90 degrees, then yea, it doesn't matter
But once I figure out all the setting and what looks good, I think it can look pretty good, due to volumetrics and lit particles everyhwere and all
It can look nice if I'm competent enough
Also I will have siltout parts with volumetric "fog"
I will make raycasts hit that as well
And light up the dust as well as the particles flying around
Might look cool, we'll see
Reflection probably isn't the fix
Having multiple rays/lights is what's meant to alleviate this
But if it becomes an issue we could look at it later
Ah I see, yes
I have 11 rays
Maybe I should add some more
Then, in theory, it shouldn't be discernible between real light bouncing and that
Yes yes I will do that
And look into light color there
Before you get anywhere close to real light bouncing with this fake method you're better off using real bounce simulations such as Enlighten or RTX
Isn't enlighten shit?
It's essentially the standard hdrp realtime gi, right?
It doesn't really do anything for me
I've researched it some time ago
People are saying it's a bad solution
There was essentially no difference when I tried it
It is the only true "realtime" GI solution there is
It works and in my experience fairly well too
Strange, as I saw a very, very minimal difference when I tried
I mean, if it works, then why am I doing this in the first place
I will try it again
Try it in a test scene first with some boxes so you can visually confirm it works, and to see what the effect is
It's not without downsides though
For being precomputed it's pretty expensive to run and not very high fidelity
It also reacts with a delay and may leave artefacts when the lighting changes, as it's a temporal effect
Since it requires lightmaps, there's a lot of types of geometry it cannot deal very well with
So it doesn't really excel at anything
But it's still effective at calculating the bounces and updating scene lighting, which are really though for other systems
Since you have no scene/environment lighting, you can't benefit from that
And since your cave is pretty sprawling, precomputing realtime lightmaps for it may be inefficient
Despite being "realtime" it does require them
That's why it didn't really work for me
No matter
Tbh creating a fake realtime gi system isn't that hard
Pretty simplistic if that's the purpose ofc
Not hard if you make it really inaccurate, that is
There's a reason why there were no realtime light bounce simulations to speak of before ray tracing hardware came around
I think unrealistic isn't necessarily shit looking
Indeed
The purpose is looking good, it's true that something that's more realistic naturally looks better
Cause we're used to it
But it's not necessarily the case
I can make wild leaps in physics and perhaps make a good looking bounce system lol
It's always a tradeoff, of development time and performance
RTX lighting took years to develop and still in most cases looks a bit crap
Even if it's technically more accurate
Even with that hardware the amount of light rays to simulate tends to be too much
So you end up with noise and then denoising which has its own problems