#🔀┃art-asset-workflow
1 messages · Page 26 of 1
Anybody have any idea why this might be happening? First time I've run into this.
In unity, the point light is situated where it is supposed to be, and the asset and the light in the viewport are positioned itentically (0, 1.32, 0). However, when building the assetbundle and uploading it to Tabletop Simulator, the light is offset significantly, as seen.
you can see this more clearly when I tilt the item:
Hey, thanks alot!
I really like some of the ideas you suggested and will definitely be giving them a try.
Thanks!
Can anyone recommend me a good tutorial for creating a stylized water shader using Shadergraph in Unity 6? I'm following this tutorial, but I'm not getting the same results https://www.youtube.com/watch?v=ILmSkM7yKD4&t=398s
"How to make your first video game" course: https://www.gamedevrocket.com/launchpad
In this unity tutorial you will learn how to make a 3D water shader using shadergraph.
Brackeys has a pretty good one, it shows how to make a basic one and explains the workflow in detail. Use that as a base and expand upon it
His tutorial is outdated as he uses PBR graph
I dont get what I'm doing wrong. I am following the tutorial with utmost accuracy, yet I am not getting the same results
Its still the same thing, just slightly different naming.
Is there a difference in the URP config between Unity 5 & Unity 6?
There are some differences, but most of them are just background optimizations and layout changes
What results are you getting?
Do you mind screenshotting?
These are the parameters
This is what the result should be in the tutorial at the stage I'm at
And this is what it looks like when I move the camera close
Shadergraph as well for reference if needed
have you played arouond with the water depths parameter?
also make your puddle deeper and see what happens
How do fps games and other games such as Lethal Company deal with view model arms that have many different items/weapons that can be equipped? Mostly from an animation perspective but also from a modularity standpoint. I just don't see a way where the animation graph is a viable option for these types of games with a lot of weapons that the player can hold, and every weapon has different animations.
@raven elm #1390346776804069396 would be the place to ask
Its likely an issue with the depth sampling. Change the sampling type and see if that fixes is
Why not? Mecanim is really powerful, and ive been using it for complex statemachines since forever
Where can I change depth sampling?
The "Scene depth" node in your shadergraph, try changing it, go through all the options and see if that fixes it
I'm new to unity so I'm just trying to get the feel of how studios do these kinds of things. I'm coming from Roblox, where animations are played dynamically in scripts, without graphs.
I guess it's not that difficult with sub state machines, but I feel like it still will get messy when I have about 100 items that the player can hold.
LC has very few unique item animations, maybe like 5 off the top of my head so they don't need any particularly clever animator strategy
The main thing of course is that the animator doesn't need to be aware of every item or be responsible for swapping them, whatever item is parented to the hand bone (or dedicated item bone) would be handled by code in a generic manner
Yeah, I will go for the approach of preset poses that most items will use. But at the same time many of the items will have custom animations and custom functionality. So a nice and organized way of doing it would be beneficial. I looked at the gdc Firewatch conference talk that they did on their animations, and it looked like a big mess.
Then use IK.
I have considered using IK, but I don't feel like it gives me the artistic control that I want.
Have a few distinct poses (e.g: long thick object, spherical object, cuboid, etc...) and then use IK to mold the initial animation base to fit all the unique items
pure IK or pure baked animation arent good enough on their own, a mix of artistic control and interactability is the go to
that is an interesting idea, I might have to look in to that
also, since my game is multiplayer. Some animations might look weird on the outside but good on the view model. Is the recommended approach to use animation overrides to fix that? I tried using the Playables API, but I can't think of a good way to synchronize that across the network
That one certainly is a mess, considering their workflow and circumstances
A lot of overcomplicated and hacky techniques in general
But that's expected when they're in a rush, don't start by knowing exactly what kind of animations and blending they will need, and are only learning how to do all that as they go
Mecanim was being developed at the time, so apparently they were doing workarounds for missing features and then once released implementing those features on the fly elsewhere
well yeah, I guess that talk isn't the best reference since that game was made 10 years ago.
More an example of what real development might look like even if it's not ideal, and what shapes mecanim can take when you're thinking on your feet
More high level than practical advice but you want to know the needs of your animation states, blending and modularity beforehand as accurately as possible, so you can plan and find the right strategies
yeah, this is basically what I've spent some time doing. I'm just trying to make everything modular so that future updates are as easy as possible. Especially since I am a solo dev.
The animator is powerful because it can make a lot of animation and transitions automated and contextual using few parameters, to take that responsibility off your code having to micromanage those
I guess it's just difficult to grasp when I have worked with both Roblox and Godot before, both using a more dynamic approach using scripting instead of graphs. But I guess the Unity approach is more tailored toward artists having the possibility to tweak and manage animations in the game. Having the animations separated from the code is something new to me.
The animator is a state machine, the type of which you will almost inevitably end up creating with code if it's responsible for animation playback
Animation clips without an FSM is very messy
The graph view has advantages but some disadvantages too
As long as they don't need unique animator logic, you can just use overrides for unique weapon animations.
But if you're used to doing that in code, it may be tempting to build another FSM to control the animator, but that's redundant and mostly steers you away from leveraging animator's features, turning it into a burden instead
yeah, I feel like making an FSM would be more difficult with multiplayer aswell. Since with mecanim, it just kind of works out of the box with multiplayer.
I thought about having just a generic sub state machine that has some preset animation clips that can be overridden by different items. And if I need custom animator functionality for some weapons, then I'll just create separate sub state machines.
I just don't know how good that would work in practice
For having a lot of items with different animations, the best approach usually would depend on how much the animations can be re-used
most of the time it'd just be enough with a generic hold animation and an offset on the item
Animator override controllers are great, which I think Pinballkitty means by overrides
The only drawback is that there's no blending between different override controllers if that's needed
Override layers can do blending but they must all exist in the same animator
by blending do you mean transitions between two clips or a constant blending factor?
Layers are separate state machines (unless synced) which you can blend between by a weight variable
ah alright, well thanks for the help guys. I think I have a good understanding on how I can approach this now!
this model is from from 3ds max but when i export the fbx to unity the left sleeve, bow and the right eye are transparent
i searched up online it was because of the normals but when i flip them the light works the opposite way because it's flipped
does anyeone know how to fix this? the right model has the flipped normals and the other doesn't
pls ping me
not a 3d model guy, but it kinda sounds like most normals were right originally and it was just some areas that had them flipped?
How do i import metal maps from substance painter into unity
i mean where do i place the texture there is no metallic map in unity
i am using hdrp
Why does the texture appear warped like this?
The geometry is essentially a few planes rotated differently with no changes to the UV
We’d need some inspector screenshots of the mesh, material, etc
yeah i only flipped those
and it still didn't work 💔
i fixed it by changing the way i flipeed the original model
Remove the height map
That did it, thanks
Also, I posted a question in the terrain3d section. Would appreciate it if anyone has a solution for it as well
would it be more efficient to just add a solidify modifier to these wingflaps or turn on double-sided faces?
the wings are planes atm
3260 tris. there can be 100 enemies on screen roughly, although these guys take up 2x1 tiles so there could only ever be 50 of them
dont need perfect lighting on them, just assume the backfaces will be seen in a minimal capacity and dont want that visual fuckery to happen if i can avoid it
solidify modifiers bump it to almost 4k
Try it both ways and profile is the only real way to figure out. I would be surprised if there is a huge difference
fair.
maybe an odd question about blender -> unity but i'm trying to understand. I kept reading that you can't translate bone constraints to unity... so I just finished a rig where I have 5 bones per wing side to give it a kind of wavy animation (sorta like a manta ray would flap it's edges). i set it up so the front two bones of each wing have a copy rotation on the remaining bones. 3 rotate together and 2 rotate the other direction to get this effect.
when i keyframed all the bones i noticed theres no curves being generated for the bones using copy rotation constraints, but it animates just fine in unity. is there an explanation for that? also I only had to edit the curves for one bone on each side to smooth the animation out, but they still seemed to respect the copy rotation constraint even in unity with no additional effort. animation wasn't baked
to illustrate: these are the front two bones with curves and keyframes each moving inverse of the other (i only move each bone by itself and the rest are either mirrored or move using the copy rotate)
rear 3 bones graph editor combined
Drag & drop in hierarchy
the wig to the bone transform
Thanks
Is your exporter baking the transform keyframes?
Didn’t think of that maybe? I am using a unity export addon
It doesn’t seem to have any such options tho
I mean, you can test really easily if there is a constraint in unity by disabling the animator and manually moving one of the bones to see if the other moves
But as far as I know unity can only import a very small subset of maya constraints
It’s either that or it’s embedded in the keyframe data somehow but not shown having its own unique curve in the graph editor.
I’m still keying the bones even if they appear to have no curve info
Might be interpolating on import between those rotations
How about in the unity animator graph?
Not sure I only spot checked it to make sure it would work as expected I’m not at my desk atm
Maybe someone from here can help me
#🤯┃augmented-reality message
I'm looking for some help with Color mapping issues with shaders
My guess is that the script you used to export handled the baking.
hello, I am new at this and I want to know something. I don't know why but when I do a simple structure in Blender using extrude and pass it to Unity, this structure change and it is different, inside the structure there is nothing,
I will be happy if someone helps me, thanks!!
The polygons likely are there but you're viewing the culled backface
You can enable backface culling in blender to view the geometry how unity views it
Or enable both face rendering on the material in unity, though the lighting will kinda suck that way
Important to learn what face normals are and how to check and modify them in blender
In order to check normal direction in blender you can select a specific drop down menu in the top right of your viewport while in edit mode. The selection noted in the image I provided displays your face normals. These will be displayed as a line projecting from the face. To edit normals, select your desired face and press Alt+N
I've got a number of models that I'm exporting as FBX files, I want to write some vertex data onto the models, but Unity only seems to read the Vertex Color data, even though I've written a float value to the vertices
Does unity only accept vertex data of specific types? perhaps it doesnt like a float value
I've even tried writing a float[3] and a vector to the vertices, none of those show up
I'm suspecting that its not a problem of type, but more that Unity might not be reading anything that isnt specifically a Position/Normal/UV/Color
https://docs.unity3d.com/6000.1/Documentation/Manual/mesh-vertex-data.html haha! so it turns out theres space for additional vertex UV coordinates in a mesh, so as long as I never need to use those channels for anything (at least for now they arent being used), that does allow me to store extra data like {123, 0, 0} {0, 42.5, 0} into the meshes
topology before/after. It's my first time doing actual retopology. Any massive red flags?
(Facial expression meshes turned off)
anyone know of any free "stylised"/retro gun sfx packs?
by stylised I mean not realistic, I don't know what to call them
Hey! Just need some quick help with the editor. I have an animated sprite where the frames are all individual .png images with names in ascending order. How do I import all of these at once as a single animated sprite into Unity?
unity doesn't really have animated sprites
there's sprites and there's animations that are composed of said sprites
just import all the images, then in unity, select all of them, create animation
This isn't quite doing what I wanted. Is there a way to generate a spritesheet or something
I know Unity has spritesheets, I've used it before, but I want to take these individual images and just have one spat out
you would do that with some other program, not unity
unity isn't an image editor
unity doesn't have spritesheets, it supports spritesheets
Is there a way to override roughness/metalness of a surface with custom maps using the URP Decal Projector? I can only see these options
dude im having a really bad headchache with my tank model
unity keep saying that a polygon ist self insterting
i tried cleaned in every mode i can
and even decimate the geometry like a crazy
but unity keeps deleting the same faces
even when y fixed topology
Triangulate should do it
Blender constraints to unity pt 2... so what is the deal here? i didnt bake the animation
damped track constraint in theory shouldn't translate
lemme try with a straight export without my addon
standard export also seems to carry it over. is it just a perk of unity 6+?
What deal are you referring to?
was under the impression blender bone constraints dont translate to unity
They do but they won't be constraints after exporting anymore, just keyframes
oh. well thats fine
And also the result of the constraint can be unpredictable, because many constraints depend on the previous transforms in blender, which are different when evaluated during the export process
For example in your video the damp is perfectly looping in blender, because blender can preserve the tail's last position when it starts the animation again
But in unity and in the fbx it loops imperfectly, because the constraint's start position was different as it didn't know where it ends up at the end of the loop
You can use constraints as you're mindful of the limitations
Mainly that it only results in keyframes, and real dynamic constraints must be recreated after importing
And that if you have looping motion you need to ensure the constraints loop too
May be more elegant solutions but a workaround I used was to play all looping animations two times, and cut the first loop out in Unity's animation import settings
i'm just trying to make it easier to animate the tail whipping around
This method seems easy enough
IK doesn't seem like that good of a fit for such physical motion
Unless you move that dynamically with constraints too and face the same issue
its not IK. damped track chain
Ah, misread FK as IK
What you have is FK
Not a big deal to export with an extra loop to ensure the start pose matches the end pose of the loop
But maybe an alternative to that is to copy the end of the loop to the start to force the constrained transforms into the correct starting position to skip having to cut off the dummy loop
thats how i've been doing all my looping animations
with some curve editing to smooth it out
some animatinos just have keyframes out of bounds as support
ah yeah now i see issues. the tail is not wagging symmetrically and its cutting off the animation frames
just gonna fk it, this is too much of a headache lol
not intended ^
Isn't the solution/workaround that I explained really simple tho
its not just the frame cutting off its like drifting more and more
just gonna keep it simple. think a spline IK might work tho?
That doesn't sound simpler
Regardless of how much "drift" there is, the offset should be precisely the same at the start and end if you have multiple loops end to end
in all the time i've been fussing about trying to make this easier i could've just bent the tail a few times lol
overthinking everything again. do that a lot
If it doesn't take much effort to do the desired kind of waving by hand, then it does be quite simple
Making the keyframes cyclic and offsetting them also does wonders
But last time I was doing a tail it was no more complex than adding the damped chain, then for export duplicate the loop, and for import adjust the clip start frame to exclude the extra loop
Nothing more to it
yeah its just 5 bones i can create the whip cycle once and copy it
with an ik paddle to support direction and height. thats the whole rig and really all i need it to do
or was the original. someone made me aware of the damped track tail thing and its cool in blender but really hard to control. spent hours just locking it to an axis
like i said, in the time that took i could've just made the tail whip cycle one time and just copied it around like i did the first version of this
other thing that was annoying was trying to remove rotations from bones... all the damped bones were not really aligning to their oriignal position unless i mashed the alt+R button
I don't fully grasp those either, somewhat mindbending that they're working off of what blender is previewing rather than what's actually on the timeline
Or so I assumed
But in FBX export phase it's linearly following the timeline
With separate actions I'm not sure
In any case loop duplication and helper poses were useful, and something I often end up using anyway when exporting to Unity
Still somewhat arcane to me when some actions just refuse to export or import at all, so I like to add redundancy
yeah not throwing shade at your method, wrote it down for future reference if it makes sense to use it. so appreciate the input! always learn something from y ou
wonder why the 3d world hasn't adopted a w3c type of standard for animation/rendering standards.
web dev was equally annoying with all the different standards every browser had
way back when
My methods are mostly what worked for me in my use cases, so only makes sense to consider all and choose what seems to be the best fit
The reason probably is that engines for the longest time were in-house or rigidly licensed so studios had their own particular workflows with no need for flexibility, and totally custom formats too
yeah so were browsers (proprietary code)
but i guess the argument there is a whole lot more people need accessibility to the internet than better animation standards. 🙄
After lease and hobbyist engines became more common, you basically had to live with the formats that existed and were somewhat supported by that point
Creating new formats with the intention of them becoming a standard is a big undertaking, and the needs of 3D were changing pretty quickly
yeah makes sense. even with the w3c there were points in time where only chrome had certain css selectors or the like
its never a perfect solution and requires the coordination of EVERYONE in the industry
cough btw hey unity ppl add nth-child to uitoolkit thx
And very little incentive to spend that effort, unless you can get some advantage by doing that
Which is why there's such battle over new image formats on the web
And I suspect it's either politics like that or licensing again that's hindering the new 3D format standard
USD is the most promising candidate, but adoption is slow
that is partly why the web is just becoming such a pile of shit in general
idk how old you are, but there was like a happy wonderful decade where the internet was this wonderful place
minus the issues with web dev
just as a user though
I think it has gone in (rather long) waves, predatory and aggressive practices gradually bloom until competitors who mop up all the crap get the spotlight
that'd probably be the .com boom
Proprietary 3D programs and other asset creation software are so expensive and honestly pretty painful to use, so I hope that kind of wave will crash there
Blender already took a leap out into full viability, might take another into dominance
blender was one of the earliest open source projects, and really was excited to see how long it managed to stick around.
bunch of open source software has retained it's communities since the most early days of the web
GIMP for sure, i forget if krita was one of those
Gimp still needs to take the leap to become viable, UX-wise like blender did
its always been kinda temu photoshop. maybe not as dedicated of contributors as blender had
I don't know much but the devs may be protective of how the program works too, so not holding a lot of hope
Game engines are hard to use in the long run without an image editor that's capable of
- technical multi-channel editing and alpha channel saving
- precise image compositing and arraigning
It's a hurdle having to choose between gimp and photoshop
Authoring of textures, sprite sheets and particles or non-color maps for shaders puts pretty specific requirements on the software
the hurdle is really just the cost for a lot of people. i knew gimp was inferior but it was always free. photoshop used to be prohibitively expensive when i was a kid
i've heard krita is pretty good for that stuff?
also free
Good for drawing/painting, not great for anything else
Gimp is plenty capable, but the UX is demoralizing for newbies and a constant hindrance for pros
its gotten worse. i remember it being pretty intuitive 20 years ago
haven't used it in a long time, came back and everything felt clunky
could just be the hiatus but i still feel like by now it'd feel fluid again
Iirc 20 years ago it was demoralizing enough that everything was a separate window, for no apparent UX reason
different strokes i guess. i liked that aspect
the path tool is so different now and doesn't feel as intuitive as it was
Photoshop is the "best" option but the cost can be a problem when doing game development at scale in some way
yeah adobe has not gotten better about any of that
everythings like super expensive license subscriptions now
Like, the class might have licenses but it's not very feasible to tell the students to pay up or pirate it if they want to keep using the workflows on their own that they learned
and they always have to split their software up into giant suites where you're missing half the features if you dont pay for 5 different programs
there used to be a educational license at some point wasnt there?
free for class use
Maybe, but won't help them at home or after the course
Don't really want to grow another PS-dependent generation either, which is really the point of those educational licenses, and which piracy supports too
A lot of students don't want that themselves either
heh
And the way Unity is set up, you really need various 2D and 3D programs externally
Luckily there's been some changes for the better there, like the texture import swizzle feature and VFX toolbox
how long you been doing unity? i know its been around a while now
First started getting into it in 2020
Why is the cube transparent
i don't think it is? you can't see the grid lines under it
it just looks like it's rendering behind the clouds
Or that clouds are rendering in front of it
#1390346776804069396 is the place to ask
sir thank you a lot
sorry for answering latter
i was so focused that i forget to respond and that was nasty
thank you for the help
keep like that fella
anyone know why the icons look fine in the editor but god awful in the game simulator? i suspected a lot of things but nothing i have tried seem to work, i want them to look as they look in the editor
You mean the device simulator in the editor?
What resolution does the device set to render?
Does the issue happen in the game view as well?
in the game view it happens, i havent tried on an actual device
i meant in the game tab sorry, it happens in the simulator as well, but when i look the scene that is being simulated or in the game they looks perfect
What screen resolution do you have set in the game view?
oh i have free aspect there
so whats the better way to do this? i have 3 LODs for each elite enemy. the shader i created has a slot for "high poly normals" which is just the normal map without the sculpt details since the high poly version (LOD0) has the fully geometry. i am just creating a material variant that checks "Use High Poly Normals" and uses the alternate normal map.
is this the right way to handle this? the high poly only appears in a specific place (a journal UI showing all the enemies you've encountered and stats for each, so an intentional toggle is fine). the LOD1 is for cutscene close-ups and LOD2 is the gameplay low poly model. they use the detailed normal map since i have that boolean keyword unchecked.
is the material variant sharing those inputs or should i just make a whole ass separate material for the high poly and cut out the "high poly" normal slot?
edit: alternatively i could've just toggled the boolean in the engine instead of having a separate material... i guess. can you toggle keywords at runtime? i imagine it should just be a plain boolean for this use case?
If you're using an actual preprocessing keyword, then there would be 2x shader variants with and without the keyword enabled.
any idea wnat it could be if i use free aspect?
So it does happen in a build..? I thought it was only happening in the game view/tab and device simulator..?
Can you be more clear about this please?
As for the solution, maybe try resetting it to full screen mode to see if the issue happens then.
sorry i might have been unclear, i have not tried it in a build, i am using the game view/tab and the device simulator
Okay, then the mentioned settings are irrelevant. What is relevant is the settings in the game tab/device simulator.
i will test them on the built game if its just the simulator that is wrong is not a big deal
i tried it on phone and they look fine, at least i cant tell if they are pixelated in a phone, gonna try it on windows now
i think they just look much much better on build its whatever i dont mind simulator looking worse i was just scared they were going to look awful on the actual game
they do look much better than in the game tab probably good nuf, no idea why they loose that much crispiness tho
Ok. now I'm getting more and more confused.
First you say i have not tried it in a build, but then you say i think they just look much much better on build and i cant tell if they are pixelated in a phone..?
So which one is it? Did you test in a build or not?
they do look much better than in the game tab
So the issue does not happen in game tab???
i had not tried them in a build at that moment, but i built them to try them a moment ago, so yeah it was true i had not tested them at the moment of the first message but i tried them in a build now
If it's difficult to explain, take screenshots of the WHOLE editor so that I can see the context of the images.
in game tab the issue is still very noticeable, in the build is still not super crispy but its much much better
Then it must be the game view/tab settings as I mentioned many times by now. Or at least, that's the first thing that you should check/confirm.
beside those options are there another place where i can see more?
No. I assumed that you have changed any of these settings.
no i left them as they are now
In this case, the issue might not be just with the game tab.
- Can you share your camera settings?
- Are you doing anything special in terms of rendering? Like rendering to a render texture?
still super pixelated icons, they look super good in the scene tho
And can you take these 2 screenshots again but with the whole editor visible?
no its just a image that loads an icon
here is my camera
and i dont know if the last image i sent is nuf? or do you need the whole screen?
Yes. It would help to see the whole picture
game or scene tab, or both?=
this one is the scene
this one is in the game tab
this is in the built game i think it looks better
Comparing the images now, they look similar
yah they do, the one from the actual game build look much better
the editor picture is much smaller tho, thats why they look the same than the game tab, the bame tab is using my whole window, the editor window is at most half
I think what you did earlier is zoom in on one of the sprites in the scene view. Obviously, the more space it occupies on the screen, the more details would be seen
yah i did zoom in on the first one
so i guess its a non issue in the end haha, in the actual game it looks fine i think
thanks and sorry for making you come with me in this wild-goose chase
any way to import the pill positions but not all the meshes? From blender. Currently I just assign them all to Pill.001 but i don't want to import the rest. Or just have them import as all referencing one mesh.
- In Blender create Linked Duplicates to begin with
- In Blender link object data after duplicating but before exporting
Or
- In Unity set all pills meshes in the scene to use the same mesh reference through Mesh Filter component
- In Unity set all pill prefabs in the scene to reference the same pill prefab
These two won't prevent from the duplicate meshes existing in the mesh asset, but if the duplicates aren't used they'll be excluded from build
yea like I mentioned i assign all the pills to pill.001 but i just didnt want to clutter up my interface with a crap ton of meshes i don't use. I'll try linking them thank you.
I have applied a decal over a large surface which has resulted in the texture itself looking blurred out. I tried upscaling the texture to 4x it's previous resolution, yet it didn't change anything. Anyone know how I can make the decal texture sharp?
What's the texture like exactly?
I think that's not the whole decal we're seeing
But it could be
Sorry its taking too long for me to reply, theres a power cutoff in my area so I have to keep my devices on airplane mode to conserve battery.
(NOW THE MAIN ANSWER)
It's a standard png texture with alpha transparency with a resolution of 3000x2500 pixels.
It's being projected on a very large plane( essentially a football field)
Smoothness is the inverse of roughness in the alpha channel,while metalness goes in the RGB channels
like that?
If you are exporting textures from painter then there is a present for Unity URP as well as HDRP for both metallic & specular workflows
when i put in to the material i have it but i need to make it in shader graphp cun i need to add paint mask and i dont see here metalic input
and if i exported ass unity present i have this and i dont know where to put this light green one
Is there a specific reason you're making your own shader for this other than some color adjustments?
Green should be the mask map
It contains metallic, smoothness and ao
So you don't have to sample that many textures for performance reasons
You output red channel to metallic, green to ambient occlusion, blue to detail and alpha to smoothness
i have shader that can change color of the paint only in my object but it dont work on new version cuz its old one or i just do something wrong so i try to do it in shader graph
You can change the tint on any shader (if that's what you're doing)
I believe all standard shaders have albedo tint. This one you screenshot doesn't do anything special that the standard lit shader doesn't do
how i can use it? i mean im new i unity and is there any chance that u can help me make it? im so lost rn
i mean i know i can change tint but it change for all, and i want to chancge only paint color
like that
Do you want to change the actual color or just tint it? If first, you should have a black and white version if the texture and multiply by color
And see how to connect mask map
If tint is fine for you, you can use any standard shader (like hdrp lit) and use the tint under the texture field
Can we switch to a voice chat? It's hard for me to explain in writing exactly what I want to do.
No, I'm at work
I understand, would you have a moment after work?
u can see that if i chancge it here its all red so i try to make it in shader graph to add this paint mask what i made before and its working as u can see above but i dont know how to add this metalic and roughness
It needs to be a submesh with own material, or separate object or decal.
but it need to be as one material one object one mesh
Submesh is the same object, it just has separate material. If you want to keep single material then recolor texture itself.
but i need to have have option to change color for multiple object. like i have 5 of this and i want to have 2 reds, 2 green and 1 orange without textruing it one by one
texture can be swapped
but if you want indivisual control over the material, make it a submesh.
i have it!
but now i dont knwo how to add bump mask in shader graph if is it even posivvle
posible
Color masking in a shader is a valid thing to do but nothing here really makes sense
i needed to make new project
something was wrong in old one and nothing was working
Setting everything that's metallic to be fully smooth and fully black is what the nodes there do, and that was the issue you were having
can i now add here the heigh map?
Make a thread in #1390346776804069396
if i want to have uv for light map i need to unvrap again in blender? i knwo it can overlap so light map uv can be difrent than normal uv map? cuz my normal map have some elements that are overlaping
No, you can simply check "generate lightmap UVs" from mesh import settings
Does a single AssetPostprocessor get constructed for each individual asset being imported?
I want to do the following:
- In
OnPostprocessGameObjectWithUserProperties, attach a component for each object with a specific piece of user data - In
OnPostprocessModel, configure these components in a way that depends on how many components I added
I already have this working fine, but I'm curious if I could get away with storing data in my AssetPostprocessor class
e.g. to count how many components I added
It looks like a new one gets constructed for each asset
Does the buildings fit to our overall look of the game? Ignore the mountain on the right for the time being.
We would be very happy to receive some feedback regarding the art! ♥️
#1180170818983051344 is for feedback
@misty lantern posted there, my bad
Hello everyone 🙂 I would need your awesome help to save my day.
so this is what I managed to get to so far in replicating my blender import to have the same lighting but im just missing something (im a unity noobie so excuse me) What do you guys think I would need to realistically add to get to replicate it perfectly. Any and all help is welcome
For "perfect" you need the same HDRi sky texture, and the same or similar HDR tonemapping
oh okay i never encountered these 2 yet
right now i managed to get it to this point
its pretty damn close, but i still have a few things that are off, like the sides of my continents are black instead of being more white-ish
Blender's HDRis are somewhere in the program's files, which you'll need as a cubemap texture into a sky material
Blender's tonemapping is in Render properties > Color Management > View Transform
Unity's tonemapping is a part of its#💥┃post-processing system
yes I have managed to find that and there are already 3 basic presets
its this if im not mistaken right?
"Blender's HDRis are somewhere in the program's files, which you'll need as a cubemap texture into a sky material"
this i don't really understand, definitely the lack of experience on my side speaks here
Your blender viewport is in material preview mode, so it uses one of built-in HDRis for lighting instead of whatever you have in scene
yess precisely
i was always wondering if there was a direct way to use it or like replicate it since its there by default, or am i dreaming too much ahaha
You can find the texture in blender's files, import into unity as cubemap, make a sky texture into a sky material, assign to scene sky and generate lighting (without realtime GI or baked GI)
can I ask a question about blender in here? tried asking in blender discord got completely ignored no help what so ever.
Hey, is there a way to make it so that textures are imported with correct max size set automatically?
It's really annoying having to go through all my textures all the time and set the res manually, I sometimes even forget to do so leading to issues.
I tried playing the the NPO2 setting and uncapping the resolution by setting it to 16k, but then all my textures were just set to 16k :/
Please @ me and thanks in advance!
you usually have to ask multiple times and bump the question till somebody notices
Oh, so turns out max size is just that, max size 🤦♂️
I think what helps the most is becoming a bit hyper-aware of which assets are NPoT to begin with, as those are something you want to avoid
Not being able to compress NPoT textures is a big downside, and they are harder to use in some contexts like effects as the aspect ratio of the texture is arbitrary
P much all of the ones I make are, so that wasn't the issue.
I for some reason thought that setting the max size would actually upscale the texture so I was looking for a solution :/
The best NPoT scaling option varies per texture, if it's applicable at all
With sprites it's not as important as you can use Sprite Atlas which compiles them to a PoT texture automatically
I see, but is it fine to just set the max texture size to 16k and import all my textures, regardless if they are 128x128, 2048x2048 or anything, and just let the unity handle it?
From what I've seen on the forums, this should be fine but idk.
As for scaling, since I am working with textures for materials, I decided to go with ToLarger as it seems like a safer option.
Yes, below the largest size that Unity can handle, the setting is a tool for you to limit texture size if you wish
Conventionally textures for materials should be square, or at least PoT
I see, that makes my life 10x easer, thanks!
Yes, i always keep them that way.
I'm trying to apply this decal texture using URP Decal Projector over a large surface, but as I increase the scale, it starts getting blurry around the edges. I tried using a higher resolution texture, but the problem still persists. Is there a way to apply this decal texture over a large surface without making it look pixelated/blurry?
This is how it looks when I increase the scale
You mean "when zooming in"?
Nope, as I increase the width & height of the decal, it starts getting blurry
How big is the texture?
4096x1840
Sorry forgot you gave the details I asked for before
Naturally whether you stretch a texture (or any image) or look at it up closer, the relative resolution will drop
Upscaling likely fails to produce results if the texture import defines a max resolution
And if it's Non-Power of Two size, it would get resized blurring it further
Yeah, upscaling is probably not adding any detail.
Is there another solution to this? I can't apply this in any texturing program as the ground itself is using a tileable texture
Getting a good high res texture would be the solution.
So, like twice the size of what I currently have?
You can increase Max Size liberally, and set NPoT resizing to None
You said you tied using a higher resolution texture, but where did you get it from?
The texture itself could be blurry too, especially if you've been resizing it outside of unity
The one I'm currently using is an upscaled version of the original texture, which I got from Freepik
I used AI to upscale it
Where can I see this option
And the original texture res is? The upscaler probably didn't do a good job.
Texture import settings
It's half of what it currently is
I might just remake the whole thing in Photoshop I guess
First you have to confirm Unity is not resampling and downscaling the texture at import
Otherwise none of your solutions will have any effect
What does "Virtual Texture Only" option do?
Open it in photoshop and zoom in the same way as you do in unity. Does it look blurry there as well?
It's related to virtual texturing, as the name implies
But that's not something unity yet supports properly afaik
Not sure if this is the correct place to ask, but what's the correct baking order?
I have lights, reflections and APV.
Please @ me and thanks in advance!
#1390346776804069396 is the correct place to ask
There's no "baking order" though, Generate Lighting generates everything that's set to generate
Only reflection probes can be baked separately from other types of baked lighting, but they're also baked if everything else is generated
I see, good to know, thanks!
@raven elm Anyway once you know what import settings do and they aren't simply nullifying all the changes you're making to the texture externally by shrinking it down, you will know if the high resolution is enough to solve your problem and if the texture memory cost is acceptable
Then you can also consider other options, mainly alpha clipping
https://wiki.cdot.senecapolytechnic.ca/wiki/TeamDS#Examples_from_Valve
Or if the color is overlayed in some way rather than alpha blended, it'll mix in more gracefully and the blurriness won't be so jarring
It'll be fuzzy more like actual paint on a grassy surface would be
Alpha clipping will make edges of transparency sharp, but still produce some jaggies on angles if you look close enough to run out of resolution
If you get into shaders and SDF textures, the sky's the limit for improving the rendering using various techniques
Like Valve's decals mentioned in the article
Nothing I did is working out. I'm completely out of ideas. My goal is to essentially apply the texture over a large field, be it via decals or any other method. If anyone has any suggestions I would highly appreciate it
How do games like FIFA do surface decals?
Did you check if it's a texture issue?
Open it in any image editor and zoom in the same way as in the editor.
Yup, it was blurry at start, but then I recreated the texture using shapes in Photoshop, and the issue still persists
What resolution is it now?
It's jus that the size of the decal is too large
And are you looking at a png? Or vector graphics in photoshop.
I retraced shapes over the PNG & then exported the output
And did you try opening the output and zooming in?
Can you take a screenshot with a similar view you had in the editor?
It's the same
Open the texture in an image editor/viewer, zoom in the same way and take a screenshot...
Is that the png?
Yes
Okay. Try using a point filter on the texture.
What did you do precisely?
Did you find and change the texture import settings
If this is viewed at 100% resolution, it is so blurry to begin with that nothing you do in Unity can un-blur it
Like Spazi suggested, an sdf shader would probably be the best option since you have shapes and text. For just the rectangles, just point filtering could do wonders. You could also simply model the decal as a mesh.
Hey all, i'm making some promo art for my asset store submission, but i feel like it's hard to keep it fun all while talking about all the technical stuff. I detailed integration in the documentation, but i'd like to know if this is enough info or if i should add more into these covers.
What's an SDF shader?
Signed distance field; it is how textmeshpro renders sharp text, for instance.
It is very googleable.
Explained in the links I posted to you
Alternatively just cut it up into multiple textures. This is a lot of area for one texture to cover.
Or use alpha testing instead of blending
How do I change the material a terrain is using? Its locked out and I can't find any way to switch it.
I am trying to replace it with a terrain material derivative that uses toon shading instead of real shading on its light/shadow, but I can't change the shader material
Oh setting Inspector to Debug will let you access and change the Material field directly
The correct way is in the last tab of the terrain script inspector
This default material is part of Unity's default assets, which you can't write to. You'll need to create a new material for this.
Theres a lot of wasted geometry in the region below that would never be visible in a game, would it at all be possible in Unity do something where the mesh could be clipped, removing most of those points (which I expect could increase performance)
As I know where every piece is going to be instantiated, that makes it possible for me to know at what height the piece is going to be underground (ie I know where those black lines will be on the piece before it exists), when I load the mesh I could immediately clip the unused geometry from it
Not worth the effort
I take unity already helps performance when it comes to mesh data that isnt visible
If part of the mesh is visible it'll render the whole thing
But we're talking, what? A couple dozen vertices per pillar?
They will be higher quality models later on, but when I'm at that point I'll have worked on other ways to improve performance
but the moment theres a bit of a performance hit that isnt a big deal, but I thought it could be nice to speed things up just a little
What are possible reasons why my Render Texture turns dark dark brown when the camera capturing it is in Orthographic mode?
The small preview isn't dark brown, but the resulting Game view is
The terrain seems to go brown in Orthographic mode the larger it's size is, but only in Game view, and only in Orthographic?
What are possible reasons that could be causing this?
There are no custom assets in the scene, its just a stock terrain, stock camera
Im notcing even the grey default square is going brown, its not the terrain, its something to do with the camera im guessing?
The problem changes when I adjust the size of the unity viewport? That's extremely wtf, I have no idea why changing the size of the viewport would change how the assets are being rendered
Am I missing something about orthographic cameras that would cause this? I can't find any answers on the net why this would be happening or how to prevent it, it seems like no one else has this issue? So it should be fixable?
oh I found the problem
I forgot I had a sobel filter edge detection in the render pipeline, somehow this was the culprit
it was me all along
Two reasons why it isn't worth doing
With opaque geometry, rendering something that's obstructed is typically skipped in fragment stage
Another is that loading the mesh for drawing is generally more expensive than the drawing itself, which is why gpu instancing can be so cheap as it can draw it many times after loading it just once
To know what to "clip" would require loading the mesh anyway, and analyzing and modifying it in a way that's easily more expensive than simply drawing it would be
Guess it makes sense if the sobel filter depends on screen space distance but isn't set up to scale with changing screen size
Tough to find considering, my suggestion eventually would've been to repro in a new project, which would've revealed a clue to where the issue was
yeah i'm going to have to re-write or find a new edge shader for my needs, working on that now atm
If this isn't the right place can someone point me in the right direction. I'm, a beginner (1st full year on and off with Unity). What am I doing wrong?
I keep getting an error for 'hips' even though there's only one hip bone. I've triple checked all my bones to make sure they were correct and not duplicated.
#🏃┃animation is the right place but the number of Humanoid users who stick around to help each other is less than single digits
How do I add more Renderer pipelines to a camera?
I need one specific camera to have its own renderer pipeline because I need JUST that ONE camera to be low resolution, and do shader post processing on its output Depth and Normals
which you can't do with a render texture because RTs dont contain normals
I can't fnd any way to achieve this look while not completely downsizing my entire game and applying these shader effects to EVERYTHING, I want to apply them to JUST specific assets, and they only work if the render size is very small
which are problems I haven't been able to figure out how to overcome
like the outline needs 160x90 size to work with to give 1px, you can't feed it a res-sized up res 1600x900, it won't outline right, it HAS to be a small resolution
I've tried render textures, materials on a raw image, URP full screen render passes, so far nothing has worked to get me that look
oh I found it, you add them through here
Does anyone know what outline algorithm produces outlines of EXACTLY 1px?
So far I've tried robert's cross, laplacian, sobel, but all of them seem to at minimum do pixel borders of 2px
There are lots of good edge detection algorithm pages online, but for some reason they keep producing 2px+ borders, so I am also starting to wonder if some part of my code is wrong and its not the detection itself at fault
You should ask these in #1390346776804069396 instead
Sure thing
where can i start to look for avatar changing in real time? (3D project)
lets say "hairstyle" , theres no way we will store a full avatar for every variations right?
like if theres 10 hairstyle, 10 face configuration, no way i would save 100 avatar prefabs right? i will only change the part (hair,face...etc)
where can i start for researching?
Not sure how many comprehensive resources there are for the topic, but examining specific games that do 3D character customization and searching for technical breakdowns may be a good way
There are dozens of different techniques that can be utilized, and each game uses them a bit differently depending on the costumes' requirements
It's also kind of a chicken-or-egg problem because your costumes and body variance determines the tech you need, but the tech you have determines what kind of costumes and body variance are possible
But generally all of the techniques are working to improve your ability to swap parts of the character, and to prevent conflicts
The basic task is to separate the character to swappable parts, such as your face model and the hair model in a way that lets them follow the same rig's weight mapping seamlessly
ty , i will start from here first 👍
Though making the parts seamless when they need to connect has some technical challenges on its own, most of the work in a customization system will be with dealing with the edge cases like:
https://simonschreibt.de/gat/mafia-ii-hat-vs-hair/
https://simonschreibt.de/gat/the-high-heel-problem/
https://www.youtube.com/watch?v=e3VkIGcBVO4 Does anyone have any ideas on how I can achieve an LED like effect on a surface? My goal is to have a PNG ad strip I made scroll over time with an LED effect on top of it
We are specialists in creating LED advertisements and this video showcases a selection of a wide ranging commercial LED ads we've produced over the past few years ranging from simple logos and branding to complex activations and events. We have created content for almost all major Australian stadiums and other local and international venues whic...
So far I have a basic Lit material with the texture assigned to emissive & bloom applied in post processing, but it doesn't look realistic enough
Which render pipeline?
More of a #1390346776804069396 question
Don't rely on automatic weights?
Doesn't seem like a unity question
What are possible reasons why these two new properties don't show up in inspector? 
I just added them myself, Im expecting it to look something like this, this is a totally different shader I am trying to copy elements from, but when I add it to my own shader, nothing shows up in inspector. Is there an extra step?
Ah I found the problem
there was a custom editor defined at the bottom of my version of the shader that was blocking the fields
Good but that's a Rendering topic again!
What constitutes an art asset workflow question?
This wasn't like shader code, it was just the inspector
Art asset integration into Unity, so nearly exclusively about importing and setting up art from external sources
Oh okay
Shader channel was merged into Rendering forum, and even if shader inspector code isn't shader code per se, it's categorically the same
Understanding of custom shaders is needed to solve that issue so that's who you want to see your question
Hey, how do I make sprite sheets not clog my search results with hundreds of sprites?
Simplest thing to do is use the slider there to change to list view, which is much more compact
Additionally you should use precise naming conventions and folder structures so you can include/exclude them as search filters
The project search has probably many features that'll help
But most likely you'll want to focus on what makes narrowing down the search easier for you, rather than hide assets from searches
In production projects asset filenames tend to be full of prefixes and suffixes to give the most precise information about their use, and to allow them to be filtered in searches
Hey, I was wondering if builtin urp materials support multiple uv maps?
We overlap uv maps a lot to save up space and lower the texture resolution, however, that introduces major issue with AO.
I was wondering if we could somehow utilize a different UV map for AO?
URP/Lit doesn't offer this functionality. It'd require a custom shader.
I see.
Is that even a good idea or are there better solutions?
This is just something that came to my mind.
Seems like a good idea
In fact, Unity offers a built in tool to generate a separate UV set for lightmapping which would otherwise run into the same issue very often
Thanks for the response.
Would be neat if builtin ao offered this aswell, it's the only other builtin map that requires separate uv afaik.
Gday everyone, I wanted to try something different than pixel art for once so I decided to practice normal art, I wanted to make some game art but I wanted to know what resolution is best for this and also how to slice the sprite sheets on higher resolution , like if there’s a difference. and do u all have any tips on this?
I'd figure out your target resolution and do some mockups based on that to get a feel for the sizes you'll need.
so i have this model of a room here, everything works fine, i can bake lighting and it works, but if i cut this hole into a wall for a doorway, suddenly the light baking doesnt work at all
and it just does this
how do i fix that
ok i kinda fixed it by just remodelling the shape
reason was probably your UVs
did you unwrap them properly
i did download this tileset but dont get it how to slice it because everything has diffirent shape and size any advice?
when i slice the column with reference of single grass block (circled in red) i got gap (circled in blue) . advice?
Slice using which method? Where is the sprite sheet from?
Can you be specific
It's a separate end piece, you can make it part of the upper tile, then it will be a single larger end piece.
you mean its unrelated to column
End piece tile is the one you mark in the RuleTile to not have anything after it
It's a 32x32 tileset
You are presumably slicing using automatic
They don't have to be uniform as well, as long is anchor set to the correct place
for an end piece it's fine to not be uniform.
are you sure
This is end piece from right and left as well
Well it's for a 32x32 grid but not all tiles are 32x32
For eample the trees
It should only conform to other 32x32 tiles on the bottom where it connects
32x32 looks like this
These would both be valid options for a 32x32 grid
probably move one vertex to the left
Tiles are allowed to extend outside grid cells in Unity
i just affraid to tile collider
These don't conform to the same grid as other tiles
If you want to work off of a 32x32 grid, it's better to start by slicing the whole sheet by grid size, then modifying
That said not all tiles on the sheet conform either, which is likely the author's mistake
There's many ways to slice a tileset, but important to note that Unity's rule tile tool example is designed to work with just one particular type, if you intend to use that
i guess there is no way working with this tile without broking tile palate
Tile's collider type can be set to "Grid" so the collision shape will be the collision cell itself
If it's "Sprite" the sprite slice's physics shape will be used, which you can modify to be any shape
Helpfully Custom Physics Shape is per sprite, and tile Collider Type is per tile even if the tile re-uses the sprite
So you have a lot of control
What does that mean?
Is there any way to make texturing in blender less painful or is it just a do it over and over until it’s easy thing bc I hate the blender texturing stuff
A color other than red?
that's the default viewport material, I haven't textured it yet, but thanks for the comment
"Texturing" includes dozens of different techniques and workflows, so I'd say yes you have more or less painful options available
The painful ones are not guaranteed to get less painful with time or experience
Depends what you're talking about
As you know, before adding tile sets to the tile palette, we cut everything into equal sizes to avoid problems. But in this tile set, everything has different size . Most likely, the tile palette will get messed up.
If I am making a 2d game and decide to go for top down pixel art it will be really hard for me to use Non-pixel art for my ui and items right ? (I have zero design knowledge)
I think the palette window is meant to get its grid size from the palette prefab
But it seems it might not
Not sure why, but at least it's just a visual issue in the palette window and won't affect your real tilemap
Can't think of any reason why it might be hard
UI is usually rendered separately
My concern is a high detailed sword in ui might look really bad if game is pixelated, for example there a loot on ground when I hover on it brings up popup with a high detailed sword, I wonder of it will blend well ?
Maybe, maybe not
That's a question of opinion
i tried every shape there is no equal way to slice neither big chunk or column
I don't know what kind of "equal" slicing you are expecting
It's a 32x32 grid, but some tiles have overspill which can be included in the tile
Or make into separate outer border tiles which would've been done in engines that don't support toiles going outside grid cells
Start by slicing 32x32 using the "by size" option
isnt they have to repeat eachother after slice what ever i do some parts got lost
When I say slice 32x32 using by size option this is what I mean
It makes the correct grid automatically
you picked a rather annoying tileset but it works
It doesn't have every type of tile and has a lot of redundant repetitions
Grassy corner step pieces are off by one pixel as I mentioned which can be corrected by moving the slices
But they are missing the top grass so the grasses have to be their own tiles unless you edit the tile sheet
That is not the same sprite sheet
what you used
The one you had here
You moved it and cropped it
It's not on the same 32x32 grid anymore
It's not even divisible by that tile size so you lose the rightmost and bottom row slices
And you cut off more than half the tiles in the process so I'm not sure how you intended to use it as a tileset
many parts clash with eachother
Tree is not for slicing, should be treated as a sprite or single tile. Also might want to move it out from sprite sheet.
That sprite sheet was not packed for application ready use, just thrown together.
There is no clash
are you think they are 4 diffirent sprite or 1 ?
Works fine, though inner corners require overlapping
Why the lime green bush is offset half a tile I don't have an explanation for
There's more than one mistake on this sheet
so you just pressed 32x32 and nothing bad happened
Yes
Besides these being offset
One of the log tops is spilling over as well, but it's a duplicate
i dont get it what are those exactly i will let hem exits but not going to use them
"stair" corner piece, so grass on top and also grass next to it
oh
Without it you get a gap
how you understand what is it order ☠️
I am very clever
Preferably wouldn't go through the trouble however, rather rearrange the tiles on the palette so it makes more intuitive sense or set up rule tiles to avoid that entirely
You still have the tiles with erroneous offsets there, so some of the pixels are now on unique tiles that shouldn't be there
Yes, like I keep mentioning the author of the tilemap didn't place them all correctly, which is why the stair pieces have one pixel offset, and one of the logs has too
Which is why you need to move them about one pixel
And delete the problematic log tiles entirely because they're duplicates
If you want the lime green bush to work, move its tiles half a tile up or down, remove the newly empty slices and add one for the top or bottom whichever needed
Since these aren't from your own scripts, they're internal to the editor
Restart the editor if they seem to break something
hey it worked
thank you
how you managed to connect them
You paint using the tiles in the middle of your selected group of tiles
How do you rename terrain layers?
XY problem, what I actually want is I need a way to get a unique data identifier at X location on the terrain, and I figure the easiest way to do that is poll what terrain is painted there
But I thats hard to do when they all have default names instead of descriptive names
the first two have the same name for example and I need unique identifiers
oh they exist as a weird nothing asset in a random folder 👀
I setup a default TextureImporter. It's working fine but I found that all the sprites I imported have the same name, the name of the original sprite whose import setting I used to create the preset. Is there a way for the preset to name new imported sprites based on the name of the file that's being imported?
nvm, it fixed itself for some reason after I deleted all assets and imported again
quick question! do i have to delete decal projector materials when I am assigning a new material to said projector? I know this is how materials are normally handled like this to prevent memory leaks. im using Unity 6, URP.
"Assigning" how exactly?
so, from my understanding when you set a material on a meshrenderer you have to make sure to delete the previously instanced material that the renderer was using prior (please correct me if i am wrong). I was wondering if the same logic applies when setting a new material on a DecalProjector? I hope this clarifies my question a bit more?
assigning via code, my bad I should have specified
Not unless the Material material is a newly created material instance
Where does material come from in this case?
pretty sure it is the unity asset, not an instance of it (was getting errors for trying to call Destroy since it wasn't instanced)
does projector.material not create instances like meshrenderer.material?
Modifying the Renderer.material is a shorthand to using new Material
If you don't need a material instance to modify you'd use GetMaterials() SetMaterials()
And if you do need to modify the material in an instanced way I'd use new keyword instead to be clearer about what the intent is
Renderer.material and Renderer.sharedMaterial are a bit too inexact to be useful for anything important in my opinion
I feel like I've assigned sharedmaterials using .material without it creating an instance but I'm not sure
Documentation suggests .materials returns instantiated materials and .sharedMaterials returns shared materials, but the method pages seem to imply they return any materials
So it seems better that when a material instance is needed, the code explicitly creates one, keeps track of it and disposes of it when it won't be used further
that makes sense. and the projector works in the same manner too? it doesn't have the Get/SetMaterials function nor does it have a SharedMaterials property. when i interact with the projector's .material property, i get errors that suggest that i am modifying actual material asset instead of an instance of it. this makes it seem like the shorthand .material instancing that is used by other renderers isn't being used for decal projectors.
hold on, im gonna try messing with my shader and see if changing the properties to instanced properties fixes the issue. might be how i have it set up that is fucking with it
was a last ditch try, but didn't work
If the material of the component is marked as (instance) in the inspector after you've changed it, that means it's an instance
Documentation doesn't mention that the decal projector does anything with material instances so I assume it doesn't instantiate any materials automatically
DecalProjector does not inherit from Renderer so it doesn't have the same methods
Alright so if I have to make normal maps for base texture atlases and some objects don't need normals and some do (we're going for weird stepped normals);
to avoid making multiple materials, should I make masks where I'll want the proper height maps, and make the rest of the atlas grey before converting the heights to normal?
does this seem like a logical workflow?
I feel like context may be missing
What "proper" height maps?
I meant to say where I want to apply normals
In my mind when you are compiling a texture atlas, each object has whatever textures they need in their spot in the respective atlas
That means a neutral normal map / neutral height map if they don't need normal or height information
Not sure where masking would come into play
No no
the textures have no normals
only base map atlases
I now need to make normals for each and they're already in an atlas
but some of the objects need normals and some don't
If you want to avoid materials, then the normal maps must be atlased the same way as base maps
Materials that don't need a normal map would only have neutral normal light blue in their spot on the atlas
that's what I mean
I'm making the height maps to convert to normals.
So I'd make a mask around the areas where I don't want normals and just make them grey
before converting to normal
I was just trying to confirm if that workflow seemed reasonable
I guess, I don't know how you're compiling the atlas
Typically it's procedural so there's no need for masking
As long as the objects without the need for the kind of maps have correct neutral data on the atlas
yes
For converting height to normal it only matters that it's solid flat I believe
If this stuff is changing in your project, it sounds it might be too early to be atlasing
well it's already done, textures are already atlased
And need to be again if they change at all
I know, I know
but the texture atlases are quite full, at least most of them so it's fine
Avoiding materials also is not worth doing prematurely
Or possibly at all if you have SRP batching
yeah I told them that too
it would be easier for me if I could just make the normals for the whole atlas and they don't need to apply it for materials that don't require them
Gotta measure that performance empirically
well I gotta do it now so
also, I still need to place filter masks on the atlases anyway unfortunately because of how the textures were made and how the normals need to look
What does that kind of "masking" involve practically?
ahh
ok so
they want normals that are not detailed, but just add the basic geometry.
So if you think of brick tiles, they would only want the normals between the bricks
but the textures aren't really made for this, so I have to do some value stepping, thresholds etc
to eliminate all the mids, and leave only the darkest darks
but different parts of the atlas have different brightnesses etc. so I really have to make maks where I step the values on just those parts (or edit the brightness or whatever) etc
So you have the base map atlas, but no tools to generate another one with the same coordinates for a different purpose?
Meaning you're left to paint over it by hand if you want to exclude any specific objects
yes, bascially
I would literally just paint gray on the parts I don't want
unrelated to all the step filters etc
Right
Or white or black, probably doesn't matter
yeah I was thinking whether it makes a difference or not
doesn't matter in any case
so anway, all things considered, that seem like a reasonable way to do this?
probably more like the only way lol
I don't see an issue with it, even if it's a bit backwards to have to do that manually
I know man, but I can't help that
If the atlasing is done simply by shared UV space, I'd rather look into making a baking setup for them in blender so it's possible to easily generate those masks and more if needed
But I'd only recommend that if that's a familiar process
it isn't but I'll mention it if they wanna look into that
Hi guys
yeah that was my mistake i was operating off an assumption which led me on a wild ass goose chase. your help was very much appreciated as always.
Hi, i am going to start making backgrounds for my game, I am using pixel art and the backgrounds need to be 1920 x 1080 to fit my camera and obviously drawing to that size is not easy, I just want to know how other people approach issues like this, should I just make a small cnavas and then scale it up in unity, I have the pixel perfect camera
I'm not an artist so maybe I'm missing something but it's not obvious to me why it's not easy. What's the problem you're having with it?
how do i slice those uneven sprite evently so that my animation doesn't look janky
Slice by cell size or count
Sprite sheets should be authored on a grid so they can be sliced with stable pivot points
Or if they're not, pivot point information needs to exist somewhere for you to manually assign it per slice
they are assinged to a pivot point, but that was in the editing software
What software is that?
aesprite
How are you "assigning them to a pivot point" in it?
the sprites are in 32x32, so i put them in the center of 4 squares like this
each square is 16x16
So there is no "pivot point"
Not that it matters much as Aseprite Importer doesn't support Aseprite's pivot point format
But they also aren't on a consistent grid either
If you were to slice them to a 32x32 grid, even with 16 pixels Y padding between rows, some 32 pixel columns would be next to each other, some not
Because some frames take more space than 32 pixels, you'd prefer to place them on a 64x64 grid
aight
You can also just manually place pivot points on some consistent point, such as a specific foot.
And when using Aseprite it's not necessary to export a sprite sheet, but an animation that gets converted into a sprite sheet
hi might be a dumb question but I want a large camera size for my game, and I am using pixel art so my idea is to use a canvas a bit bigger then 1980 x 1060 to allow the pixels to look good and give the player a look of room to look. so my question is to have a bigger camera size and have my pixels all be one to one accross all assets i have to draw bigger background right?
Can you give an example image of the scene you're working in? What you're asking is still not quite clear
You can have a pixel perfect camera reference resolution of 1920x1080 to be played on a 1920x1080 size window for a one to one pixel resolution, if that's what you're asking
sorry what im asking is do I have to make a big canvas size for my backgrounds if im doing parralax backgrounds and want a big camera size
like is there a better way to do it as the solution I have found is to make my pixel art with a canvas of 1980 x 1060 which is obviously gigantic, it would work, but would take lots of time
You're mixing unrelated things so it's still unclear what the actual problem is. If you're asking if you need to draw the backgrounds using 1980x1060 image size in whatever software you use to draw them, then no. If the pixel art's pixels correspond to 5x5 on screen pixels then you can draw them as 396x212 images and so on
i know UI billboarding can be done with simple transform scripts
but the best way of making a billboard UI image (image that will always facing camera in any angles) , is by shader right?
even this seemed a bit overkill with just one demand?
Parallax layers will have your camera more or less space on that layer
Less on background layers, more on foreground
Details on foreground layers tend to be repeated, but any detail can be repeated (and should, whenever possible) so you rarely have to give the same attention to every pixel
Tilemapping when authoring art or in engine is one way to save effort by repetition
UI is typically facing the camera to begin with
If you have a world space Canvas, you can rotate the Rect Transform of an image with a script, or the Canvas itself to avoid repeatedly regenerating the whole Canvas geometry
It's not overkill, but the cost depends on how many of them you need
Regenerating a Canvas with many elements in it is typically the expensive operation
normally my billboard world object consists of one button with simple structure like this , but it will be put into many kinds of objects
so u cant tell whether they will have 10? 20? 50? or 100? , ofc they will not be active until certain condition is met
i would say maximum under 30 at same time? 30 seemed a bit lower than expected
One thing you can do is rotate all objects in a loop from one component, but it might not make any perceptible difference even if all 100 had their own components
I'd prefer not trying to combine them into just one Canvas, but you'll only know the difference if you test
i just made a shader instead , since they are more performant with large amount of objects right?
it might be 30 now, but i cant tell i have other more UI world objects in future, so i played safe
As far as I know Canvas shader graphs don't support vertex stage shaders
So you might lose some canvas features if you use a non-canvas shader
lol from what i found most of the shader tutorials are for non-UI stuff , they must require vertice
i guess i gonna do it by scripts
made some fluorescent lights 100% in Blender as a challenge. I normally use substance painter like a normal person, but I think I got pretty close to hl2-esq PBR with just Blender and some addons. The workflow seems to scale better at a large quantity of assets, compared to substance painter, which was the initial motivation of doing it.
got a question about workflow - i'm coming from the programming side of things, if I wanted to make a handgun model for use with VR would the whole gun be modeled in a single blend file? or do the interactable components - slide, hammer, trigger, magazine, need to be modeled separately and "assmebled" in unity?
It would be all part of one FBX file, you wouldn't have to do them separately no.
Of course they would be separate parts within the FBX file, so that you could move them.
and by "parts" those are just separate "objects"/"entities" in blender
Yep
like i don't need to do anything special to declare them as such
great, thanks
3d is calling me. i can feel it.
!collab
:loudspeaker: Collaborating and Job Posting
We do not accept job or collab posts on Discord.
Please, use Discussions to promote yourself as job-seeking, advertise commercial job offers, or look for non-commercial projects to participate in:
• ** Collaboration & Jobs**
Is there a 'general' channel or what is the next closest thing to that concept? I just need to get some eyes on something in an extremely casual way
I dont want to start a whole new thread just for someone who is around to get some eyes on this 
i guess unity-talk maybe?
hello, for creating level should i make it in blender or directly in unity, my biggest concern when making it using blender is the integration for example let's say that the 3d artist placing bunch of vegetation if we export it to the FBX the tree will be count as different game obj instead of prefab, and if i try to make it directly in unity let's say that i hire some 3d artist and they never touch unity it will definitely slow down the process since they need to learn
This is why 3D artist and level artist are two different positions. You can be one without any knowledge of the other, or you could know both. Some studios, like Naughty Dog, do their levels inside Maya, with very custom pipelines for bringing that into engine. However in the vast majority of cases the levels would be done in engine
ohh thanks, this was for my project with friends, the one who making the level never touch unity before but have experience in 3d using blender
also it is possible to instance things in the 3d package and have them point to 1 mesh in unity
but yeah i would make parts to build a level in 3d pacakge and do the layout in unity
yeah, my suggestion to my other friends is to make editor tools, to get vec3/pos of the mesh, and then replace those mesh with the new prefab, but yeah meaning that it introduce extra step
idk what i am missing, does texture size node not work as supposed with _MainTex like sample texture 2d does?
Do you have an example of it not working in a particular situation / in a particular way?
Hey guys does any of you have experience with selling 3D game assets?
my bad for not elaborating, I was expecting Texture Size node to benefit from MainTex similar to how Sample Texture 2D does. and not having to assign the texture
I think it does work
However there's a kind of a delay for the material to update after any changes to the shader
Going into play mode seems to allow the sprite renderer to correctly set the _MainTex in a way that's readable to the Texture Size node
I believe because texture size is not directly accessible to the GPU so there must be extra steps internally to deliver that information
going to play mode does seem to fix the issue. quite unconventional
that does make sense. thanks for your help
Does anyone here have experience with blender and unity?
note that VRChatSDK is not supported here, z
Hey guys! I'm super new to unity and I need a bit of guidance. So I made a bunch of 24x24 sprites and everything looks fine when I throw them in the sprite editor. But when I try to make an actual tile palette out of them, they look super tiny and virtually unusuable. Any help would be appreciated!
Hello I'm making a spacegame with ship size around 150m in Blender or even larger. I don't think it's
a good idea to use this scale in Unity because of floating point issues when moving the ship far away
so what is the best scale factor to use at import to scale it down to an appropriate size?
That is usually solved with floating origin (the ship stays at (0, 0,0) and the world moves instead)
But if you don't use that the scale should be such that relevant objects are at most 1000-5000 units from the origin
the problem with that is I don't have only one ship but more so could be a bit tricky and I was hoping to get around this by scaling it do you think 0.01 is to small ? because I guess at some point it's too small for the Engine to display everything correctly
Floating point precision goes both ways, so if vertex coordinates are very small the model won't render correctly
For astronomical distances you need to do smoke and mirror tricks anyway. Having more than one ship isn't necessarily a problem
so the general idea is that I have a point of interest like a Planet or sth an then you have like a space around it where you can go around. Its more like a RTS game and then you can like jump to different points of interesst so I was thinking maybe I can just move them to the origin but still with a ship of 150m I only have like 33times the ship length to work with so I was thinking of scaling it down but there must be a sweet spot in wich the Object should be scaled right ?
Yeah, small enough that when you fly around it never goes further than 1000-5000 units from the origin
So it depends on how big the space is
yeah I didn't yet decided a fixed size for the space or even if its fixed it was more like finding a good scale that gives me a lot of space and also a good range of objects being different size like a planet would be much bigger and also not to small so that the objects renders correctly but yeah like I said I don't know if 0.1 oder 0.01 is best ? like its difficult to say at best I take the smallest possible so I have enough room to work with
You'll have to experiment yourself. Scale the model down, import it in Unity, put it in a scene a few thousand units away from origin and rotate and move it around. If it looks ok then all good
ok will try it thank you
Wrong PPU
PPU should generally be the width of one tile or grid square in pixels
Oh, gotcha! So, I'm a little confused then: what should my PPU be if my tilesets already look this jank at 24x24? Should I go for something lower?
Not exactly
PPU value in import settings is determined by the pixel dimensions your sprites are authored for
It's a convention that you decide of how many pixels you mean to fit in one unit of space in the scene
If you redraw the tileset for a different size pixel grid, like 32x32, their resolution will be higher (and perhaps look more or less "jank") and so their import PPU must be 32 to not spill or cause gaps in the tile
Both those tilesets will fit their tilemap grid snugly, but the 32 pixel/PPU one will be of higher resolution and practically has smaller pixels and more of them per tile
Oh, gotcha! I'll try to mess around with the PPU some more and will report back. Thanks a ton!
It may help you understand what it's for, but generally "messing" with it won't give you exact results
Each sprite / sprite sheet that's drawn with the intention of being placed in a grid will have a specific correct PPU which is the same as the pixel width/height of a tile sprite
Which you can measure
How should I go about measuring it? Either way, should I move away from turning this sheet into a tile palette then since they'll all have a different PPU? And instead import them as individual images?
They won't have a different PPU
The PPU is the same across all your assets, for the sake of making sure each sprite pixel is as big as every pixel of every sprite
You sliced them by 24x24 grid and everything fits on that kind of grid so that's one indication the PPU should be 24
Otherwise you would select a group of pixels the size of what you want to be one "unit" or tile in your drawing program and see how big of an area that is
Most marquee tools report the selection dimensions
I'm sorry, but I'm a little confused. So if my PPU should be 24, and it is 24, then why did your earlier comment say my PPU is wrong? Is there a mismatch somewhere?
If the PPU is too large (like default 100) then sprites will be much smaller than the tilemap grid in scene and in palette
This is what appears in the first screenshot
However, if you try painting in the scene's tilemap with them (on a 1 unit by 1 unit grid) and they correctly fill the grid cells, that would suggest it's a tile palette preview issue rather than PPU issue
I did not see you mention what the PPU was, or what you changed it to and what the result was
Or if you already painted in scene using the tiles and it looked either right or not
I don't know if you specified a non 1x1x0 tile palette size, or grid component size
Ohhh, gotcha. Yeah, you're right that it looks like a tile palette preview issue rather than a PPU issue. They correctly fill the grid cells.
As for changing the PPU outside of 24, I switched it to 36 and while it worked for some assets, the rest are weirdly sectioned off and had a different sort of jank
I'm outing myself for how new I am to all this stuff, but I have no idea what those words mean. Time to do some reading up lmao
Which ones?
this part: "non 1x1x0 tile palette size, or grid component size"
Grid component has a "cell size", which determines the size and aspect ratio of the grid
Because a lot of things like physics, lighting, camera size and Pixels Per Unit values are relative to units, cell size should generally be 1 (or either width or height should be 1 if you use non-square tiles)
Just to keep things standardized and sensible (although there are some specific reasons to have non-1,1,0 grids, like quarter-tiling)
Tile palettes have the same property, which you probably want to set to the same values as your scene grid even when 1,1,0
Automatic sizing can do weird stuff sometimes, and will do if you have sprite slices of more than one size in your tilemap sprite sheet
Ohhh, gotcha! Thanks a ton for the explanation. I'll be sure to avoid the automatic sizing. I think I fixed my issue? Had a PPU thing set at 100 and changed it to 24
That's exactly what I was getting at, in so few words
not sure where to ask this, but why do my lights cause this?
does anyone know any good topdown (or angled but not isometric) pixel art environment packs (ideally a 64x64 dungeon environment)?
did you check the asset store?
I did but I couldn't find anything good
found this after 10 seconds of googling, is 64x64 and looks pretty good https://free-game-assets.itch.io/castle-platformer-2d-tileset-pixel-art
Thanks but that's a side-view, I need something more like the background in this game which is why it's difficult to find
It's also weirdly difficult for me to describe it
it's called isometric, and there's a bunch here https://itch.io/game-assets/tag-isometric/tag-tilemap
Explore game assets tagged Isometric and Tilemap on itch.io
Ah looking at it again you're right, the perspective threw me off lol
Thank you I'll take a look through these!
Does anyone know how to make unity keep the image texture look the same as it looks in blender ? Like i can take a image and put it on something in blender but the image looks completly different in unity, the scaling is way off
What do you mean?
Could you elaborate on the difference
I found out, I was trying to put a image from Google onto an object in blender and then export it to unity, but when I exported it to unity it was all stretched out and blurry but I found out how to fix it
Yeah, generally when I export to Unity from blender, I export an in FBX and set the path mode to copy
It defaults to auto
How did they achieve soft & fluffy grass like this without taxing the system too much? There's even a lot of small grass blades as it's visible in the sillhouette
I think you are talking about non-power of two auto resizing
But it helps to give some clues or even examples when you ask
probably a shader
That image is way too low quality to tell for sure, but probably a shell shader.
Anyone here working with a Blender-to-Unity pipeline and putting all of their game's models into one big .blend file, using separate collections with per-collection exporters?
I'm just looking for opinions on this type of workflow. It's working reasonably well here (I'm exporting each collection as FBX etcpp.), and it gives me the advantage of having one big 3D workspace to work with, seeing all my game's models in relation to each other, and so on.
Only issue I have is that when I tell Blender to export all collections, it'll happily generate new/changed .fbx files even for collections whose content didn't change at all, which is a bit annoying with version control.
I'm just trying to find out if what I'm doing here is for some reason insanely dumb, and I should be doing something else instead.
I'm just trying to find out if what I'm
I have been given the task to create a realistic cricket stadium with this as the reference. I have to make sure I make the field look as accurate as possible to the reference. From what I can understand, this will be a mix of custom decal textures projected on top of a base realistic tileable grass texture, but I dont understand how I will achieve it
Moreover, I'm having issues with the normal blending of decal textures
You sound much more advanced than me, I have no idea what I'm doing here tbh :P
If you're asking about per-collection exports, just create multiple collections, and in the collection tab in properties you'll be able to configure an exporter just for that collection. (It's a relatively new feature)
i guess you could also make all of it with just a shader, without needing any decals
If you want to showcase use #🏆┃daily-win or #1180170818983051344
thank you
hi, im trying to get a tileset into my game but when i try to slice it stuff nearby ends up bleeding into tiles where they shouldn't be. on the right are my settings
the original tileset was too small so i sized it up in aseprite (i think this is where the issue is coming from but im not sure)
any help is appreciated
nvm, i think i found the solution by increasing the max size
For realistic brick walls, is it best to use a flat model or a more complex one? Will good textures carry the realism far enough?
Good textures can work...if they are done in a way that allows a 3D look to them, best bet is use normal and height maps/bump maps
That by far most depends on how close the observer should be able to look at them
The second most important thing is what you consider "realistic" personally
If it needs depth up very close then in addition to geometry (typically with LODs), tessellation, parallax / parallax occlusion mapping are useful tools
But if it's not seen at an arm's length or at glancing angles in sharp lighting, these methods may not be as necessary
Question about working with artists:
What is a recommended way for artists to share models to programmers to use, while in the same team?
Like we're doing a school group project with artists, and currently our workflow is that we programmers teach the artists how to use github desktop & unity, and show them where to upload the models and stuff in the unity project itself
I'm wondering if there's a better way? some kind of tool or something that I haven't been able to find info about?
Learning Git is good and useful, but only if they'll need that skill
In practice it'll only skip the effort of importing the assets if it's up to the coders to implement them in engine anyway
If the artists will also then do tasks within Unity, then it'll be more worth the trouble to teach
But if they work functionally separately then whichever file sharing service that seems most convenient will work
We're using git as the filesharing atm
Yes
So if the artists won't need Git later in life, any cloud service would be the better option by virtue of being familiar and skipping the effort of learning/teaching
What would you recommend as an example cloud service to use?
I have at least 1 artist who's not interested in touching Unity at all (so I don't see the reason to show them how to use Git for an example)
Any that's familiar and easily available, like Google Drive or OneDrive typically
There's more specialized cloud storage services, and more binary asset focused version control software, but with them you're quickly moving away from simplicity of Drive, and from ubiquitousness of Git
And also decreasing the likelihood that the service is offered for free
There may be some "obvious" alternatives to Drive for project asset use, but someone else will have to recommend those
I've only had to choose between the two options of "teach Git / a real VCS" versus "don't spend any effort at all and just get it done with whatever cloud storage is at hand"
I think i'll probably stick with what we've done so far, because it doesn't really take that much to teach git (at least the desktop app version)
plus it's good for the artists to have an extra skill set if they wanna stick to movies or tv shows as they have been so far
Hi guys, I'm a complete noob at Unity and wanted to ask a specific question about rigging.
I want to use these free assets and rig them in a way where each part will easily be replacable with a random one, just basically mix the characters all over the place with each body part. How hard is this to do and what exactly do I need to do to achieve that? 😄 (IDK if this is the right channel to ask in)
Sprite Swap will let you do it rather easily
IIt's designed to work as a part of 2D rigging which you'll likely use to rig them
thanks, do you have any pointers, like guide or tutorial i can watch? Sorry im a complete newb xD
Both of the Tutorials im watching about 2D Rigging people are saying its obsolete now and Unity made some big changes in regards to that?

What tutorials are those
And have you looked at the official documentation
Game Dev Brotherhood: https://www.skool.com/game-dev-brotherhood/about
TIMESTAMPS
0:00 - Intro
0:22 - Step 1 (sketching)
0:33 - Step 2 (Refine)
1:14 - Step 3 (Importing)
1:42 - Free Game Dev Course
1:59 - Step 4 (Placing bones)
3:10 - Step 5 (Weights)
4:31 - Step 6 (IK Controls)
6:20 - Step 7 (Animating)
7:57 - Step 8 (Challeng...
im watching this but for some reason now around 5:20 when I'm trying to add this Effector, my bones don't work like on the video, I can't stretch it along the arm and when I 'Create Target', I can't move the arm around like the person does in the video
What's the best 2D rigging tutorial in your opinion, something up to date as well? 😄
I haven't really seen a up to date AND good one.
maybe u know a shit one thats at least up to date? 😂
Not off the top of my head
I'm still using an old version of the package, myself, for my main project.
The one you linked seems ok at covering the tool, they just are bad at rigging.
Though depending on your needs you don't necessarily need best practices.
I created the overall texture for the field using a mix of decals & tiled textures, but now I'm facing issues with creating the actual pitch. I need it to blend with the grass texture, but I can't find any such texture online. Anyone got any ideas?
I'm unable to figure out how to make the pitch decal texture blend with the ground
anyone happen to know how do i import file from blender, im having trouble. the white part are image textures
Did you also import the textures
i use copy mode
Setup materials in-engine, not in Blender.
Blender does not have Unity's shaders. It can't create or match Unity materials with full accuracy, even if some of the basics can be mapped.
"Images" are textures
The name just informs the context how it's meant to be used
enable alpha clipping on the material
ah that was it. thanks a ton
Alright, lets get into something weird. I'm making a 3D tileset, but the game I'm making is pretending to be a pre-rendered 2D game. Here's an example of what that looks like:
I want to be able to model the tiles as cube-shaped as possible, but that will cause trouble when we get to things like these stairs, that have to be seen on both sides.
This cabin is also a really good example:
I've been getting a sense of how this would work. If the face of the cube/tile is static, but the back of the tile is scaled up and/or moved around a bit, I can achieve the desired shape. My question now is, what's the best approach to doing this for an object I'm going to treat like a tile-set in Unity?
The first step would be to detect all of the nearby tile objects so that I know what the nature of it's transformation will be. I'm just not sure how that transformation should take place.
Should I scale/move the vertices along the camera-facing axis?
Should I animate the 3D cube in blender, making each possible shape it's own frame of animation that the object could be set to in unity?
Those are the two ways I think I could do it. I still don't know if there are other ways, or in any case, how I should
When you say a 3d tileset, you mean that the tiles will be 3d, or that they will be renders of 3d objects?
Theyre going to be 3D assets entirely, but I need to stretch their back half so that it can be seen by an isometric camera. On-screen they'll look like the 2D tiles I posted, but still made of 3D objects
"Preserve specular lighting"
Alpha clipping culls pixels below the alpha threshold value entirely, so it hinders semitransparency
Not necessary for this purpose
Typically you'd use opaque materials with alpha clipping when you don't need semitransparency, or transparent with no alpha clipping if you do need semitransparency
Preserve specular lighting you'd use for semitransparent glass or membrane-like materials
there’s no semitransparency here, looks like a good case for alpha clipping.
Likely a good case for opaque material as well
You rarely need alpha clipped transparent
There's no real type of orthographic projection that would match your reference
The farther side of the objects would be physically wider, and the projection would be typical orthographic
If you were to do it by transforming vertex positions instead, you'd have to scale the farther side horizontally relative to some point, like the camera, but that'd lead to a bizarre and nauseating parallax effect
Also, you notice that while the objects seem wider at the far side, the gaps between them seem narrower
That's doesn't make sense unless they are physically wider, and physically closer
true
made my first ever sprite earlier today, hoping for some feedback
this is the protagonist, I wan't to have multiple styles unlock if I get that far so some ideas for that would be cool, or just general feedback
not sure how to make it bigger for discord lol my fault
I wanted to ask a workflow question
Simple question: unpack or not unpack FBX models?
I have a humanoid character that has meshes, armature and animations inside.
But from time to time again we had to at least once change the fbx(UV, weight paint, extra bones, etc.)
I have found that dragging the character into the scene as a child of a game object. Then adding the necessary scripts, colliders and other meshes without unpacking was much easier than unpacking it at any point.
Replacing in the windows explorer was easier simple and kept project intact
How do all of you do, or how it should be done?
any resources/ tips I can follow to make anime style shaders in unity
I made this in blender
Learn about toon shading, or get an asset off the store that does anime style shading.
can anyone recommend me some lighting for games courses? Im specifically looking for urp.
try udemy
I think gamedev.tv has something
It has some on shaders, less so on lighting specifically
Like your name 
can anybody help me improve this sprite its for a game and its my first time doing pixel art please tell me what i should add/remove
I did check their site, but I couldn't find anything specific to lighting.
anyother place i should be looking?
Hey Y'all,
I'm curious if anyone has experience with vertex painting to blend between materials in Unity. I've built a few variations of the shader itself allowing the blending to happen and utilizing height maps to make a more believable transition, but I'm mostly interested in the workflow y'all have used in the past. Do you use one of the vertex painting tools that can be put in Unity (Polybrush or VertPaint for example), or are you doing your vertex painting outside of Unity in your 3D modelling tools? This is a workflow that I've never really found good reference material on, and in area I work professionally, this isn't something we would often need to do.
Image included as a reference from Dying Light: The Beast as there are multiple buildings with the exact same wear, even if base colours and whatnot vary slightly (which they appear to do). Broadly I am assuming this is done via vertex painting and blending materials as it feels wildly excessive to have this actually unwrapped and manually painted.
EDIT: Additionally if you know of an alterative way this type of effect would be achieved please feel free to shed some light on that side of things
Decals is an alternative way that Unity better supports
I'd probably do this outside of Unity. Doesn't seem very difficult to make a shader that does this.
It's not given that this is done with vertex paint. They could be using a mask map to blend between different features of the shaders. It's a way to get better resolution blends without dense geometry
That's my main issue with vertex paint. The other is that it breaks batching. Two identical static meshes with different vertex paint will not be batched together
I use polybrush + a custom shader for (mesh) terrain. For something that you can uv map, you can do it outside of unity
I had a question. Like I have a blender file with hand and gun rigged with animations of idle and walking only. I export it to unity in fbx format. But few days later I want to add another animation of running and climbing do I need to import it again? And like is there some way it replaces the current Gameobject placed in scene or like update itself automatically without need to place it again in scene and creating scripts and animation controller again?
are people using git LFS for holding your texture atlases and whatnot?
or rather, at what scale of asset development would LFS be required?
Pretty much anything beyond code or other text files.
Decided to do some optimization trickery, it aint one to one but damn will is save triangles
trying to improve the performance of my room
You can use #1180170818983051344 or #🏆┃daily-win to show progress or to ask for feedback
ahhh that makes more sense, thanks!
thanks
is SpeedTree actually still down?
I just straight up can't use the program because they've been twiddling their thumbs for a month? tf?
Ask them
I mean, asking "it it working yet" isn't going to do much besides annoy them. I emailed support for a temporary key.
More of its users there in either case
Pretty wild if an always online tool has been offline for a month
the website is up, but they aren't allowing new purchases and the website is super broken rn
so anyone that wants or need to use speedtree can't export anything
Hey guuuuys
I got a sort of finicky question about texture atlasses, if anybody here is a wizard about it
We've been using unity Sprite Atlasses for our workflow for the whole project for memory optimization, Now one of our TAs is coming back to report that we may have misunderstood something.
He says the original uncompressed, original size textures are still being included in the build, and the atlas is only for runtime RAM memory.
So if we have 3 1k textures with some dead white space in them but we made fitted sprites and atlasses those 3 into a single 1k x 1k texture, he says those original 3 1k textures are still being included in the build download size.
This runs counter to how everybody else on the team understood this to work.
- Is this true?
- If so is there some way to tell unity to discard those originals and only use the atlassed sheet it generated in the storage memory?
- Does this belong in another channel?
You could probably confirm that in a build report.
That being said, theoretically unity only includes stuff that is expected to be loaded at some point of the game. If these textures are entirely irrelevant due to their color and metadata being copied to the atlas, I think they should not be included. If something references these textures, then they wouldn't just sit on the disk, but get actually loaded into memory. Which you can also confirm via the memory profiler.
Maybe ask the person that reported it of where they got that info from and/or how they confirmed it.
I'm not exactly sure if this is the correct channel to ask in, but does anyone know why I cannot apply prefab overrides here?
Exactly the right place to ask
Model Prefabs are prefabs that are automatically generated to contain all the objects and components that are in the imported mesh format
As such they can only be controlled by the asset's import settings
Okay! How would I go about fixing this? I'm not very good with prefabs so sorry for the vague question
What is your goal in this case
Do you intend to use prefab variants here?
I'm not sure to be honest. All I really want to do is override the prefab so that every instance of the spike has a collider
Tbh I don't even really know the difference between prefab variants/model prefabs/normal prefabs 😅
I would make a "Spike" gameobject, reset its Transform, drag the spike mesh "model prefab" from Project window onto it, add and configure the collider component on the "Spike" gameobject, then drag it into Project to make it into a prefab
This way the spike mesh exists inside the Spike prefab, and you can modify or even swap the spike mesh without the Spike prefab caring about it
I would suppose the prefab with import settings is the model prefab?
Yes, the blue and grey box with the black strip of tape on top of it
Icon, I mean
Now you can modify the Spike prefab in prefab context mode and changes will be saved
If you modify the prefab in scene outside of prefab context mode, the changes will be local overrides to that instance of the prefab, and you can "apply" the overrides to make them global to the Spike prefab
Ok! I tried this and it seems like the overrides are applied instantly to the instances in the scene. They all have the collider but there are no overrides to apply. Would you say that's a good thing?
You only need to apply overrides if you want to change the scope of the override
Prefab 1 could be the parent of Prefab 2 and override Prefab 2's properties
So every instance of Prefab 1 contains a modified version of Prefab 2
But Prefab 2 if used alone will not have that override
I think I understand
Applying that override from Prefab 1 will make it a permanent feature of Prefab 2, so even if used without Prefab 1 it will have that change
Or also commonly applying per-instance overrides that exist only in scene into a prefab globally
Scene per-instance overrides are the only context where applying overrides works like "saving changes"
With nested prefab hierarchies it's not necessary to do unless you specifically want to change which prefab contains which specific changes
You could have a "Spike Platform" that contains many of your Spike prefabs, but shrunk down
In that case you don't want to apply the smaller scale override from the Spike Platform to the Spike prefab because that change is only needed by the Platform
alright, it's all a bit hard to grasp but thanks a lot!
I think there are some pretty good official tutorials, and it'll become intuitive soon when you practice it
@finite fiber Model prefabs are a bit less intuitive because they can only have unapplied overrides, as their structure is totally generated by the import settings
I find the best workflow is to let them exist in the prefab hierarchy as separately as possible, because they can change a lot if you modify the mesh or the import settings
Like, your Spike prefab can place the collider component onto the Spike model prefab and keep it as an override, but if you later, say, correct the import scale from 100,100,100 to 1,1,1, the collider's relative size will change drastically if it's on that transform
I see, thanks!
anyone here who worked with speedtree? the export is placing the pivot point at center of the tree rather than its bottom which is super annoying. is there a way to change that
nvm found it's a unity thing
Is it possible that they placed the imported files into a resource folder or some else that is always included in builds?
Yeah we found a Unity Thread that follows up on this as well. We're going to do some digging today.
It's not my area to spend too much time in, it just sounded really fishy and worth looking into before we redirect our workflow to include manually created PNG atlasses.
The Thread: https://discussions.unity.com/t/sprite-atlas-adding-on-build-size/727423/15
No, all our images are in an Art folder, not resources. But thanks!
Ive been having some trouble with importing some assets into unity and having them work with a shader I created. The shader is for leaves to create some fun gradients and applies darker colors to the bottom of the object regardless of how it's rotated. I dont have an issue with the shader, but rather some objects that I import into unity do not work with the shader at all and cause it to not function correctly
instead of the shadow color applying to the bottom, it just changes the obejct's color when it's rotated upsidown. This only happens with some trees I made in blender. when using a default cube from both unity and imported from blender however it works just as intended
I used geometry nodes to create the leaves on the trees, but some trees work just fine, but others dont. Tree trunks also have this same issue for whatever reason. I applied all transforms, set face on normals, imported into maya and exported into unity with the same issue. I can export two objects with the same exact settings in blender and a default cube will work just fine with the shader but the leaves will not
That seems to be more about Addressables than about sprite atlases per se...
My mistake! I had too many threads open.
Link updated. It's a conversation going on from 2019 to 2024 so people are still encountering it recently
Honestly it's not clear to me from the video as to what the issue is. All of the demonstrated objects see. To behave the same to me.
Sorry it might be hard to see cause of the grass but if you look the darker color is supposed to be on the bottom of the object and stays on the bottom regardless of how it's rotated
On some it works, on others they're completely light and when rotated turn completely dark and the colors do not behave like shadows
I fixed the issue though. For whatever reason applying transforms on the leaf objects completely breaks their ability to behave normally with the shader. I manually have to set their origins to 0,0,0
So which of the objects shows the correct behavior?