>>1025688
bad bake

>>1025688
>Fake Ai video
Get that Ai vid out of here, only real artist post in /3/

>>1025688
I don't get why people keep posting that. DLSS will never find its way into official Blender due to unsolvable license incompatibilities.

>>1025699
wrong
it won't get implemented due to strong lobbying against by BIG EEVEE

*Cultist general
Blender sucks

>>1025688
>Someone needs to leak that build
Literally first result on google.

https://projects.blender.org/blender/blender/pulls/153077

>>1025699
Can't it be released as a third-party add-on?

>>1025711
autodesk should release maya as a blender add-on so cultists can stop seething about it

>>1025688
does this look good in motion or does it get those edge artifacts and lighting ghosting that dlss gets in games?

>>1025739
honestly, I don't care at all if those are preset. if it's as smooth as these goyslop videos show, it's a massive stride for working within the viewport. I don't need pixel perfect shit in motion, I just want a fast live render view. the final render is what matters anyways

>>1025750
yeah the viewport is great, i was just wondering if you can conceivably use it for the final render itself. probably not for anything meaningful

File: real blenderguru.png (138.4 KB)

138.4 KB PNG

https://www.youtube.com/watch?v=kSzGkF65YVI

File: file.jpg (57.2 KB)

57.2 KB JPG

goldmine of seething retards in comments
https://www.youtube.com/watch?v=nlc44Tsd_bM

>>1025791
>NOOOOOO my heckin creative works
No one cares about "creative minds" from the third world like all of them are, 3D is oversaturated anyways and it's time to cull a big population from it with their shitty generic mediocore work. AI is the test if you get filtered or not, basically if you're low value shit.

File: 1771175576271.png (29.9 KB)

29.9 KB PNG

>>1025799
>t.

>>1025688
Dogshit. You can literally see a reduction in detail through this shitty kilabyte bitrate video. Who dont you post a 4k version at lossless quality and see how eell that goes.
"Muh shimmering hurts my eyes". No.

>>1025688
so in order to use this we have to compile it ourselves?

File: alicia pris.jpg (620.5 KB)

620.5 KB JPG

I’m trying to turn a character model I captured from a PS1 game screenshot into a proper 3D model. The original mesh had holes in the head that I repaired, but now those areas are discolored. I’m wondering if that’s because the repaired faces aren’t properly mapped on the UV layout. If so, what’s the best way to fix this kind of broken texture? I also want to rig the model with a skeleton, but I’ve been told that doing so isn’t ideal because the character isn’t in a T-pose or A-pose. How difficult or time-consuming would it be to reposition the model into a proper pose? Is it really so bad that I’d be better off continuing to spend hours trying to rip the original model correctly or even just rebuilding the model myself from scratch?

>>1025979
Some o5her models from the game are on models resource so just compare their setup to yours. As for t-posing at that low level of complexity it's a 30 minutes job max

>>1025791
clickbait nothing here is about ai

>>1025791
slowly but surely every creative will use AI
he's a pussy though, his audience is already dragging him in the comments

on a long enough timeline every artist will abandon their luddite ways

File: buggy-as-fuck-1_FAXRmvfB(1).mp4 (1000.1 KB)

1000.1 KB MP4

>>1026005
Yeah, I think Panta would be the better when it comes to "retard's first model into VRchat & Garry's mod" because I gave the model to a guy who put it through a program called "Rigify" or something and this happened though I do have some questions about that...

File: Panta error1.png (1.6 MB)

1.6 MB PNG

>>1026022
I'm trying to follow a guide to port this guy into source but I can't follow the guide because of this error It's due to a modifier I think, but that modifier is there to keep his backpack attached onto him. Is there a better way to make it attached to him?

>>1025688
I don't use denoising

>if you want weight painting to be practical, you have to go to Tools > Options > enable Auto-Normalize
I just found this out. Idk if this is a Blender specific issue but this is why 3D modeling is kinda pissing me off. Like in other artforms it's about how you USE the tools. In 3D modeling it's more about FINDING the tools. Like "Oh you didn't know about this secret hidden option that can save you 10 hours? teehee XD". I've watched like 5 rigging videos and only through some random youtube short did I find this. This isn't the first time something like this happened.

is there a way to array an object along a spline without it deforming?

>>1026038
Have you tried googling it?

>>1026034
Depends on your workflow. The armature modifier will always normalize the weights on the fly so it kinda depends on the effect you want your brush to have. With it on paiting weight will take away from other bones, otherwise it will not.

A typical use case is using auto-normalize on the basic skinning but turning it off when weight painting corrective bones, which are typically gonna have very small weights "on top" of the basic skinning.

>>1026038
Yeah, create a line mesh and deform that, then in the object tab set up vertex instancing, then parent the object you wanna "array" to the line mesh.
Then if you really wanna control it with a curve add the curve modifier to the line mesh

>>1026043
Although you may want to use a ribbon mesh instead and instance on faces because that'll give you better control over the instances' orientations

File: ivy.jpg (186.7 KB)

186.7 KB JPG

>>1026045
Ah yeah that's pretty nice. Used a plane array as the face instancing source

do you guys use bforartists? is it a meme

>>1026084
>bforartists
it has some cool things but all the tutorials are made for the default interface so it is better to use that instead.

>>1026034
every software has this exact problem. 3DSmax and Maya have an insane constraints system that effectively requires coding to get it to play nice, on top of functions requiring dummy objects to make the bones follow constraints (they can't reference another bone, you see!)
Blender's constraints are honestly the best, and even then they're not great. I'd prefer them to ever going back to Autodesk though.

>>1026084
Two reasons why I'm not interested
1) I've been using Blender for years and its UI feels intuitive to me, there is a logic to it
2) I don't like using forks that might get discontinued

>>1025688
Add primitive tool in sculpt mode is in main now, we could have it in the 5.2 beta tomorrow.
https://projects.blender.org/blender/blender/issues/149216
https://projects.blender.org/blender/blender/pulls/152856

>>1026113
Circularize operator from "Loop Tools" in main too.
https://projects.blender.org/blender/blender/pulls/151642

>>1026113
I would need to try it, I like keeping objects separate when blocking out but this looks ergonomic

File: 17721662674044063423274903642152.jpg (1.2 MB)

1.2 MB JPG

Any tehcnic one can learn or read up to do walk cycle, or idles, or even sequences, and all three inbetweened, procedurally, maybe using just simple flash/2d timeline or in http://tululoo.com/ with simple objects, events, +-/ maths

File: Screenshot 2026-02-27 120613.png (1.4 MB)

1.4 MB PNG

Is there a reason a lot of the crap I find on the internet has this sort of bone layout, and is there a way to convert it to tip to tail connections so I can use IK? I just put a new armature on it if I need to actually animate it, but I feel like theres a better way.

>>1026173
Select all, rotate locally by 90° to realign them. Show axes to find which axis to rotate around, will be the same for all bones. Ripping program merely used different axes

>>1026067
ty
>>1026041
rope faggot

>>1026201
More or less how you do it. I've had to do this a few times myself. It also helps if you use the cursor and "Selection to Cursor" on each of the bones' heads to make sure the bones are properly aligned with each other, especially where the fingers are concerned.

It's time consuming, but it's easy to get the rhythm of after a few times.

File: Sculpt Scene Project Brush.webm (964 KB)

964 KB WEBM

Sculpt: Scene Project Brush merged today!

>>1026354
https://projects.blender.org/blender/blender/pulls/136342

>>1026354
NICE.
I've been wanting something like that for awhile but didn't know there was someone was working on it.

Im trying to learn how to animated an would like some help with how workflow. I want to used the WASD navigation, set camera location, and then key frame it. I usually then move the keyframe around depending if its too fast or slow between parts. I need to get used to camera navigation if anyone has any tips, thx

I'm having a major issue with Blender. Randomly it will cause massive slow-downs on my system. Sometimes I can fix it by tabbing out of Blender but most of the time I have to restart my PC. Nothing else causes this, only Blender. I don't remember this happening a couple of months ago.
Could outdated plugins be the problem? Please help a retard out, I just want to continue modeling my shitty robot.

>>1026425
If it's really extensions you can try disabling them in turn, but I'd sooner suspect a driver update

>>1025688
ahh man I really love me some poopy fart butt smears all over my renders! gotta love that Nvidia slop!

I'm so close to figuring something out. But it feels like I'm reaching the limit of my ability to comprehend. I know how it should be done, but at the same time, I can't put the pieces together. I'm excited to show you guys, but I'm also unable to finish it. Very frustrating. If only I was a little smarter.

File: Clipboard_03-15-2026_01.jpg (22.7 KB)

22.7 KB JPG

i have no idea what i am doing

How much vram is needed to work comfortably with Blender? I'm buying a mobile workstation (post-display or leasing) and either I will get 4GB (T1000 or T2000) or 6GB(Quadro RTX3000).
I'm willing to use older version if needed.

>>1026667
4.5 was working fine for me with a 15 years old quadro 2000 so you really won't have any trouble

File: Untitled.png (186.6 KB)

186.6 KB PNG

is this good? I'm trying to make a low-poly "segmented" mesh where each limb is a separate object for easy animation without worrying about weight painting and stuff.

>>1026668
Nice thanks!
Were you making a lot of stuff? I would like to know the scope of projects so I don't end up suddenly hitting a wall.

File: Clipboard_03-16-2026_01.jpg (22.1 KB)

22.1 KB JPG

>>1026666

>>1026670
Well it's not bad.
I would only say to maybe taper his ankles, and the wrists of his hand segments.
His child bearing hips are low key giving femboy fetish. Take that as you will.

>>1025688
Fucking ew.

DLSS looks better, but the compression is still shit.

I still get excited to update blender, there's always at least one feature or improvement that's useful to me

File: HOW THE ACTUAL SHIT (1).mp4 (1 MB)

1 MB MP4

>>1025688
WHY DOES AUTOMATIC WEIGHTS CHOOSE TO DO THIS?
IT LITERALLY NEVER DID ANYTHING LIKE THIS BEFORE

>>1026711
Nevermind. Forgot to invert normals

>>1026354
Did they remove it? I don't see it on release notes for 5.2

File: _ (Unsaved) - Blender 5.2.0 Alpha 21_03_2026 10_03_15.png (193.3 KB)

193.3 KB PNG

>>1026767

Is Apple M5 Pro (20-core GPU) good enough for rendering? Or at least for doing work, but rendering on some other machine?

>>1026880
yes

File: Delaunay Triangulation 01.mp4 (2.5 MB)

2.5 MB MP4

>>1026634
I pushed through. Check this out. Delaunay triangulation. My very own custom Delaunay triangulation node. Technically, the fill curve node already has a hidden Delaunay triangulation function. So making my own is redundant. So why make it? Well... This is just the beginning. I'm trying to comprehend how voronoi meshes are made, so I can make one in 3D. One that I *control*. That's key. I don't want to just have random points scattered, and make the voronoi mesh off of those. I want to control the points, and make the voronoi mesh based on positions I choose.

Anyway, understanding 2D voronoi is a stepping stone to understanding 3D voronoi. So this is a step closer to the goal.

File: Delaunay Triangulation 02.mp4 (134.3 KB)

134.3 KB MP4

>>1026897
Here's a comparison of Blender's fill curve delaunay triangulation with my version of delaunay triangulation. Blender's version is the one with the boarder, and mine is without the boarder. They are identical. Blender version is way faster though. I mean, I thought mine was fast when I was testing only about 16 points. But it starts getting clunky around the 200-300 point range. While Blender's version can do 10,000 points in an instant.

But that's ok. I'm just happy that my version technically works. It doesn't have to be the fastest for now.

>>1026897
>>1026901
Well, if that makes you happy

>>1026904
It does. There's a purpose for this. When I reach the end goal, it will be glorious. You will see.

File: Delaunay Triangulation 03.mp4 (1.1 MB)

1.1 MB MP4

>>1026901
Creating the 3D version was easier than I thought. Though, I didn't really solve the math portion. I was just lucky that someone already created a node that finds the circumsphere of a tetrahedron. All I needed to do was plug that into what I already had and make a few adjustments to account for a 4th vertex. Now check it out.

File: Delaunay Triangulation 05.mp4 (2.5 MB)

2.5 MB MP4

>>1026946
I changed the method slightly, which sped up the node about 3x. Now it reaches about 1,000 points before it starts to chug. I do want to make it faster, because ultimately, I want it to be able to triangulate between thousands of points. Maybe if I coded this instead of used node, then it would be faster. But I don't know how to code.

Anyway, I also figured out how to convert the triangles into voronoi cells. Like actual voronoi and not dual mesh approximation. Pretty neat, right?

File: Delaunay Triangulation 06.mp4 (887.5 KB)

887.5 KB MP4

>>1027003
Having a bit of fun, and moving a point inside of the voronoi cells.

>>1026634
>>1026897
>>1026901
>>1026946
>>1027003
>>1027007
Looks great, anon. Especially the last video looks fun to play around with.

>>1027003
If you want this to be fast it needs to be coded in C with space partitioning and threading. Cool student project though.

>>1027034
Chud

>>1027008
Thanks.

>>1027034
You're probably right. I dunno. I don't know what space partitioning and threading is. But what I suspect is holding back my node from being faster, is 2 things.
1, the repeat node is just slow in general. It would struggle to repeat *any* function 1,000 times. I suspect that repeating is faster when it's actual code and not nodes.

2, When storing values, there's an option to only store selected values. So say you have 100 values you want to store, you can select to only store 20 out of those 100 values. But even though the option to do that is there, I suspect that blender nodes are still checking all 100 values, and then deciding to only change 20. And by checking all 100 values, it's still doing the work. I suspect that in coding, if you only want to store 20 of 100 values, then it will actually only check the 20 and ignore the rest.

And if my suspicions are true, then the process I'm using would speed up immensely. Because I could tell it to only check the circumcircles/circumspheres of a few triangles/tetrahedrons per iteration.
As it is now, it's checking all existing triangles/tetrahedrons per iteration.

So yeah, maybe coding would make this way faster. I dunno for sure though. Still, the "project" doesn't end here. When you see what applications I actually have in store for this, you'll be amazed. I can still demonstrate the application, even if the speed of the node limits how far I can take it. Maybe after people see what I'm creating, then they'll be motivated to develop it properly.

>>1027038
A modifier needs to be as fast as possible, if you want to use it for procedural workflows, animations etc. Though, if you intend to use it as a tool that you tweak once and then apply to your geometry, for example a remember, then longer slower processing is fine. Not ideal, of course, but you could get away with it, depending on the particular usecase you have in mind.

>>1027057
>A remember
*A remesher. Thanks autocorrect ...

File: Delaunay Triangulation 07.mp4 (3.3 MB)

3.3 MB MP4

>>1026946
Phew. It took a lot of trial and error. But I finally figured out a way to use the speedier method for 3D Delaunay triangulation. It's still not as fast as I want. But it scales up a lot better than the previous method. In the previous method, it took 5 seconds to create 200 tetrahedrons. And trying to push it to 300, it just made my PC stutter for a couple minutes before I forcibly shut down Blender.
The current method takes a little under half a second to create 300 cells.

It's kind of difficult to see the tetrahedrons when they're all packed together, so here's a video of them breaking up.

>>1027057
I want to use it for animation and procedural stuff.

File: Delaunay Triangulation 08.mp4 (3.9 MB)

3.9 MB MP4

>>1027099
Turned the 3D tetrahedrons into 3D voronoi cells. The transformation adds 50ms. So now 300 cells takes a full half second, instead of under a half second. Not too bad.

I still don't know how to make the triangulation conform to a boundary mesh. Most voronoi demonstrations show a neat little box broken up into cells. But I don't know how to make it conform to a box yet. It'll look more impressive when I figure that out.

Modeling hair is painful

>>1027123
This is neat, but how is it going to help you make anime titties?

>>1027131
The most advanced algorithm for breast bounce simulation, finite elements modeling, relies on having the breasts turned into a 3d delaunay triangulation

>>1026354
I take it this could be useful for clothing?

I discovered an error with the triangulation node. I'm not sure what's causing it. But sometimes the position of the points causes it to create more faces than intended, and the extra faces intersect. It's something I need to fix before moving on, so I probably won't have any cool triangulation updates for a while, until I identify and solve this problem. Honestly, it was a miracle that in the span of just a couple weeks, I went from unable to build the node, to having a functioning node with speed improvements, and being able to transform it into voronoi cells too. The progress has been explosive. But that kind of progress is something that I can't always expect to have. Oh well, I hope I fix the problem sooner than later.

>>1027131
Delaunay Triangulation can be used to calculate volume. Which will help with physics.

>>1027133
>finite elements modeling
I didn't know that term existed. Interesting.

Total beginner here, still learning. Can anyone help me attempt to T-Pose this model? Appending it to a scene get me 90% there, but the main body has a weird issue where it turns to a flat plane. Any advice would be appreciated.

DL: https://tstorage.info/rbw5iykgmcj7
Password is 77840

File: file.png (75.9 KB)

75.9 KB PNG

Has anyone made a plugin or a geometry nodes setup that approximate 'loop subdivision'?
I want to use it to smooth out a bunch of old game models that are already triangulated.
Normal subd looks ass on them and their topology can't just be converted to quads.

File: Delaunay Triangulation 09.mp4 (2.9 MB)

2.9 MB MP4

>>1027165
I think I identified the problem. When points are at the edge of a tetrahedron's circumsphere, it can cause the resulting new tets generated to be flat. Or "coplanar" as it's called. The flatness means that the new tet has zero volume. And that causes problems. I don't know what to do about it yet. I tried deleting the flat tets, and that did prevent slow down. But not entirely.Some odd tet generation still occurs.

Still, the triangulation is usable, so I figured I would work on the sampling the tetrahedrons. Like how to tell if points are inside specific tets. And then how to interpolate attributes based on the points position. These were fun things to work on. I think I finally comprehend how barycentric coordinates work. Maybe. I was able to do the sampling math on my own, and the numbers appear correct, as far as I can tell.

Setting up some crude hair curves, I set up a crude marcher that extrudes vertices by sampling the interpolated attributes of the curves. And the result is pic related. It's a mess, right? But you can see what I'm going for. A curve interpolator, using Delaunay triangulation as a frame. I need to work out the kinks. But this should give you an idea of the application.

File: Delaunay Triangulation 10.mp4 (2.8 MB)

2.8 MB MP4

>>1027226
My node for getting the area of a tetrahedron needed a little tweak. Fixing the error seemed to solve the bulk of the interpolation problems.
Still not perfect, but a lot better.

>>1026354
Hey that's actually pretty neat , could be useful for on-skin clothing's

>>1026666
>i have no idea what i am doing
Welcome to the club

File: Screenshot_20260405_142552.png (1.5 MB)

1.5 MB PNG

>>1025688
Help plz
I want to learn how to make and use control rigs for 3D models. Like last year I followed a tutorial but the person was clearly very novice and I encountered a lot of issues with the rig when I actually tried to use it for animation! Honestly it may have been partly my fault because I don't even know enough to use the rigs properly, but on the other hand many of the controls he set up in the tutorial just like totally broke unless the character was standing perfectly still facing +z with both feet on the ground.
Either way, can you fine folk please recommend any tutorials on making and also using pose rigs?
I want to learn how actual real pros do this stuff, if there is an actual pro tutorial or course or something that teaches this well I would love to hear of it! Any advice would be very appreciated! I want to try making animation as good as the amazing digital circus show! Their poses and animations look so natural and clean, but mine just looks like some tard mashing his keyboard on a 3D software. My poses are absolutely lifeless :(

>>1027282
That's nice, pookie bear. But you're not allowed to post stuff like that on a blue board. The mods around here might overlook a nude wireframe. But I don't think they're going to overlook two penises rubbing heads.

>>1027270
P2design channel for tidbits, art of effective rigging for the full course

>>1025979
>what’s the best way to fix this kind of broken texture?
Yed, go to UV Editing
Select the new face and move it around the texture
Sometimes you have to split that polygon and place it where it should go in the texure

In my opinion for a model this simple, you better do it from scratch.

File: uPKtIBLPSXSUcOYAmcVX_Kibali-s+Face_square.gif (2.3 MB)

2.3 MB GIF

>>1027289
Thank you! This does look like exactly what I wanted!
Has anyone here taken this course before? How was it like? the website says it's appropriate for beginners, is that true?

I'm feeling a Blender related fatigue setting in when it comes to tutorials.
So much so that I might just ask an AI for answers of Blender related questions to improve my workflow.
Not asking for generations of 3D models or textures or renderings. Just instructions on how to do that kind of stuff.

Anybody tried this? How good or bad did it work for you?

>>1027319
Why don't you ask the AI whether it'll work?

>>1027303
It's true, but motivated beginners. It starts from real basics but ramps up fast and is quite lengthy, but oh so worth it. But you could also start with one of hus simpler free courses to get a feel for it.

>>1027319
I've always searched for answers on google. Then, one day, I looked at the AI results, and it had the answer I was looking for. So I don't discount the usefulness of AI in seeking solutions to problems. Typically, google's AI is sourcing from a particular video or article anyway. So it's giving you an answer humans came up with. And if the AI's explanation doesn't make sense, you can click the source to see if it's actually relevant to your needs, or if its erroneous. You still have to use your own judgement. But it seems that AI does help the search.

>>1027323
Maybe in the future such questions will yield genuine results, but I see the point you're making.
>>1027326
Fair enough, anon.

How is open3Dlab for getting some basic models to test render settings in Blender before the sculpting and texturing begins?

What I'm currently concerned about is render quality. I handicapped myself into using Linux (legitimately not sure if that means anything in terms of improved Blender results, but then again people say Windows 11 is awful) with my AMD gaming rig so I want to find a sweet spot between efficient renders that meet a compromise between high quality and not making my PC hurt too much.
I'm not sure if Blender can meet what I am looking for.
But really this is just a hobby I am considering as a pastime when I have free time and I'm not feeling down in the dumps.

I have an AMD Ryzen 9 7900x CPU and an AMD radeon rx 9060 XT GPU.
Am I being too concerned about pushing my hardware to the limit or are my concerns genuinely justified?

>>1027328
Been rendering on my 10 years old toaster and it's fine. Rendering is about scene management, not having the fanciest machine around

>>1027329
Thanks anon!
Now I gotta break free of daydreaming.

>>1026354
damn that's sick

File: A0CB3EF2-6B00-4FC6-AEC0-B2AAC631A602.png (858.2 KB)

858.2 KB PNG

Working on making some low poly characters

File: F5971A63-2F7F-4761-A28B-8B2F24C50320.png (589 KB)

589 KB PNG

>>1027334

File: F8F93625-61E5-4F38-861A-A3D21D9B0D85.png (582.6 KB)

582.6 KB PNG

>>1027334
Low poly face controls sure are interesting. Basically built myself a switch case for color and alpha

File: blender_rM9blJhuwq.jpg (693.8 KB)

693.8 KB JPG

I've been banging my head against a wall for the past 2 hours trying to figure out how to get this god damn light to work properly.
With the mist turned on in EEVEE, the light will just disappear if it's not in in front of something like the glass panels, and I've been trying all sorts of gradients and camera distance shit to make the light disperse as it goes out instead of terminating with that giant circle of white.

>>1027345
>EEVEE
That's your first mistake.

>>1027347
I don't want to go back to slower rendering though :(

>>1027348
You will waste so much time getting EEVEE to look right for anything but the most basic scenes that just using Cycles will end up being faster despite the render times.

>>1027348
Get a better graphics card or use the CPU.

If you can't, then give up 3D.

>>1027347
Eh, if it's good enough for Milky Subway, it's good enough for /3/

>>1027345
what did you use to create the mist?

>>1027349
>Cycles will end up being faster despite the render times
How much of a difference is Cycles and Eevee when it comes to rendering?

>>1027361
Cycles is by and large a traditional CGI rendering engine that focuses on bouncing light and simulating materials in a realistic way.
Eevee is basically a video game renderer that takes a ton of shortcuts for the sake of speed, and often ends up looking unrealistic or needing a fair bit of trickery to achieve a very complex visual effect.

>>1027370
NTA, but for Cycles, is it possible to use the denoiser for animations without weird flickering/artifacting? Is there a strategy for reducing/removing the flickering while denoising?

>>1027371
idk, I haven't really had that problem.
Maybe try a different denoiser or toggle GPU denoising.
If those don't work then you might just need more samples for a chance at consistency.

>>1027361
Cycles will take a second where Cycles will take hours

File: output.webm (2.9 MB)

2.9 MB WEBM

>>1027349
Fuck it, I'll work with the limitations, there's no way it'll take longer than the straight of rendering I estimated it was taking for cycle.
>>1027353
What mist

>>1027374
You said you used mist in eevee. Its or just a principled volume material?

>>1027370
So in a strange autistic way.
Am I most likely able to configure Cycles to produce an "EEVEE" video game render effect and keep the render effect visually consistent as Cycles and Blender keep getting updated over the years?
>>1027373
How many hours are we talking depending on how complex a scene is?

>>1027373
>>1027377
Whoops, meant eevee will take seconds where cycles will take hours. How many depends on your scene complexity and features

>>1027376
I used the mist layer thing

does anyone here know anything about geo node and script? I got an add-on that can add collision, but only to a specific kind of mesh. I checked the mesh's geo nodes and its all there.
What I want to do is apply this geo nodes to whatever mesh I want so I can add collision to anything I want.
Tried to copy paste the geo nodes but something came out wrong. Figured this is way above my level.

>>1027399
Adda node modifier to the new target mesh and select the same geonode tree you used on the other

>Quad Remesher for blender is still broke, doesn't allow for vertex painting to control remesh frequency
I messaged the faggot and he got back to me on the 3rd of Feb saying it'll be fixed in a few weeks. what the fuck did I pay all that money for?

>>1027402
>paying for a Blender plugin
Skill issue.

>>1027401
I did as you say, I clearly ctrl A and copy paste everything yet the copied geonodes didnt look like the original geonodes. How is this possible, its copy pasted!!

>>1027411
2 things.
1, It sounds like you didn't follow the other anon's instructions. Because you're "copy and pasting". You shouldn't need to copy and paste. You should be able to add the node group from the modifier tab. The fact that you're copying and pasting, makes it sound like you're trying to move the node from 1 file to another file. Am I right? In that case, try using the "append" function instead. Go to File > Append. Then select the file with the node group you want to use. Navigate your way to the node trees folder, and select the node group.

2. It sounds like you're attempting to paste the node group into a different version of blender. That would explain why nodes look different when you paste. Geometry nodes updates often. So new versions might not have nodes that old versions have, and they system automatically attempts to convert the old nodes into new nodes. But if you're pasting into an older version, it won't know how to convert the new nodes, and will likely break.
If the node group has some specific function that relies on old nodes, and they're automatically converted to new nodes, then that function might break regardless.

Check which version of Blender the node group was built for. The publisher will likely tell you on his page. That should be the version of Blender it works on. If you need it to work on a newer version, then either figure out what's breaking and make the necessary fixes, or ask the publisher to update their node, or try using the last version of blender that has all of the old nodes.

>>1027412
thanks I managed to get the geonodes through modifier tab. But it just didnt work on the mesh I grafted the geonodes on. Do you want to take a look at it?

File: Untitled.jpg (21.9 KB)

21.9 KB JPG

>>1027412
oh right, the original mesh has those attribute thingys and my mesh doesnt. I see boolean and stuff there. Consider that this is a collision geonodes, those things might be important. Can you copy them?

>>1027416
>thanks I managed to get the geonodes through modifier tab
But are the nodes the same or different now? Check inside.

Can you copy them? Well... I'm not 100% sure. But I'm thinking no. Attributes in that section are typically custom made. Specific to the individual mesh. Something the node creator did to his mesh, in order to make his node work.

Because you see, in geometry nodes, sometimes you can't generate an attribute exactly as you need. So you have to create an attribute and manually set it by hand, and then go into geometry nodes and fetch the attribute you just made. For example, UVs work this way. You manually set seams on the mesh. Those seams are an attribute. And then the UV unwrapper looks at the seam attribute to figure out how to unwrap the mesh.

So the question becomes: Is the node using those attributes? If yes, then how is the node using those attributes? Again, you should check with the publisher of the node, and see if he has any notes about these attributes. I can make guesses, but I can't tell you anything with certainty. position sharp_edge, and UVMap are typically created automatically when you create a new mesh. It's strange that your mesh doesn't have those. I'm not sure what Color is doing there, but that might be connected to vertex colors.

In geometry nodes, attributes are fetched by their names. So *if* the node needs those attributes, then you can create your own. Just press the little plus button there, and create a new one with the same name, attribute, and data type exactly.

But the attribute merely existing, doesn't necessarily mean it will work. If you need to actually set attributes(like seams or something), then that's what you will have to do. Or, if the script is supposed to generate those attributes automatically, and it's failing, then you'll need to figure out why the script is failing.

It's possible that your mesh needs to be properly unwrapped.

File: bobo.jpg (91 KB)

91 KB JPG

>>1027417
alright choom guess its above my level after all.
Wanna take a look? the add-on is pretty cool and kinda fun to mess around

>>1027418
Sure. I'll look at it. I never messed around with a collision node before.

>>1027419
mediafire D0T com/file/s96gqu5qlnovqc3/soft.7z/file
BOBA

>>1027420
I opened the PDF, and it looked like too much work than I'm willing to bother with right now.
Did you open the PDF? Did you look at the instructions?

>>1027421
no you just install the addon, import the xxx.obj, press N to open the side menu, click on the xxx mesh, choose the collision target on the menu, and press start and its done.

>>1027422
the problem is the xxx.obj is a daz 3d default mesh, and the addon only works on daz3d mesh. you know daz studio right?
I want to use this addon on any meshes I want. Atleast get the geonodes to apply on any meshes

>>1027422
>>1027423
The PDF says it needs to be installed on Blender 4.2LTS specifically. Did you do that? By the sounds of it, you didn't, because you said nodes were changing when you tried to copy and paste.

>>1027424
yeah Im using blender 4.2LTS. Nodes were changing cause I went to the geonode editor and ctrl A and copy paste them. Later on i used the ctrl + L link method to copy the modifier directly.

>>1027424
oh it works now? huh? I went into the geo node editor and changed the collision target there instead of using the modifier tab. Need to subdiv the mesh a bit too

>>1027424
fuckin hell its lagging as fk and I only subdiv the mesh up to 1 million vertices(2 millions triangles)
blender just couldnt handle it at all I have a not bad CPU.
this is a bust. when will technology improves, sigh. such a simple soft body tech and yet...

>>1027427
I got it working with a different mesh too. By selecting "target is planar surface" in the modifier tab. That seems to be the trick.
And yeah, after subdividing the mesh to 3.6 millioin tris, it slows down a lot.
Collision is never simple. Sadly. I went through that phase of thinking collision should be easy. But I learned that in order to make it work universally, every point needs to know where every other point is. And that leads to exponential costs. If you only have an 8 point cube, then 8x8=64. So you need to perform 64 operations to keep all the points aware of each other and not overlapping. That's easy. But when your mesh has 1 million points, then it becomes 1m x 1m. Which is literally a trillion. And no machine is performing a trillion operations in a timely fashion.

So people think of ways to do fewer operations and still get decent results. The node you're playing with is one such attempt. But even as it tries to be fast, you're coming to find that it's not fast enough to get the level of detail you're looking for. This is the typical of most collision systems.

What animators do to get around this, is perform simulations on much lighter meshes, and then transfer the motion onto denser meshes. Creating a lower poly orb in the shape of a breast, simulating the collision on that orb, then transferring the motion of the simulation to the actual breast.

>>1027431
Who or what is performing calculation in real world? each person has likes trillions of atom and they are constantly self collide.

>>1027428
Check the subdiv modifier and see if it tells you it disabled GPU subdiv (which will make it slow for high poly counts) and possibly the reason (which will usually be other modifiers, custom normals or a a driect dependency on geometry (like parenting something to a face or vertex).

You can then deal with it by either removing with the relevant dependencies or setting the viewport subdivs to 0 (you can do it globally in the render simplify tab), which will massively accelerate everything and is verx handy while you're animating, and obviously doesn't affect the actual render.

>>1027431
Ah yeah, perform collision on a simpler mesh. Heck, it can be a second instance of the same mesh only with subdiv turned down or off

Whenever I try to render a scene that I've created, the render window remains blank for several minutes before Blender stops responding. I'm not sure if this is because my computer can't handle scene. The scene has a total of 24,725 vertices. Here are my computer's specifications:
>i5-3470 3.20 GHz
>RAM: 8 GB
>NVIDIA GeForce GT 710
Please don't laugh.

>>1027447
God, and I thought my GPU card was underpowered.

I think that might actually be the problem. Did you try CPU render? Or are you attempting to use EEVEE?

>>1027448
>Or are you attempting to use EEVEE?
Unfortunately, that was the problem. I tried using EEVEE because I thought it would render more quickly, but I didn't realize that it uses the GPU exclusively. After tweaking the settings to make it a little faster, I found that rendering with Cycles works relatively well.
Sorry for the stupid question.

>>1027447
That sounds like eevee shader compilation. Are you doing anything fancy with materials? Also what Blender version?

>>1027447
>>1027450
idk how common knowledge this is but if you render while in wireframe or solid view in the viewport it helps performance

>>1027451
Blender 4.4. Perhaps I should switch to an older version, probably Blender 2.8.
>>1027452
>buy better hardware i guess.
I intend to do it, but not right now.

it is a truth universally acknowledged, that blender can do anything maya can do, so like how do u do bifrost rigging in blender? anyone got a tut? example:

https://www.youtube.com/watch?v=kvaHoIbVCeg

>>1027473
they have that in the workboard since forever but a guy made an addon for free
https://www.youtube.com/watch?v=6IsHUzLsUOY
i think the animation team is working on layers rn

>>1027475
https://projects.blender.org/blender/blender/issues/154504

>>1025688
Hows texture painting in 5.0+?
I'm currently on 4.3 and it's really not very good. No layers, brush engine is meh, -extremely- laggy.
I mostly just do cavity masking stuff then take it straight to image software to paint there, but there's always a disconnect b/w the model and what I'm working on there.
Are any add-ons worth using for this purpose? Or should I try and learn a real texture painting software?

>>1027484
ucupaint?

>>1027484
>Hows texture painting in 5.0+?
It hasn't changed yet, it will change in the next 2 years.

>>1027486
>Hows texture painting in 5.0+?
Ucupaint is the way to go, the developpers are working on the 3d brush, it is something that the guy from ucupaint hasn't implemented, and after that they'll work on layers.
https://www.youtube.com/watch?v=i89WJ-ap4C4

File: what the fuck blender.png (104.5 KB)

104.5 KB PNG

How do I de-duplicate my drivers? I honestly have no idea what the fuck is going on here— In reality there's only maybe 20 or so? but it's listing easily over 100 with dupes

im on 3.5.1 (yea yeah i know i should update)

File: file.png (259 KB)

259 KB PNG

why's that one box gray and why does it shift over when I click another into orange and then click away!!!!

>>1027505
showing active quad?

File: image_2026-04-12_205209040.png (453 KB)

453 KB PNG

>be me
>been fucking around in blender for years at a hobbyist level
>always wanted to make a [spoiler]tanboard[/spoiler] video game
>Sees a magnificent model deform showcase with every major muscle group accounted for in topology
>no tutorial fml.
>FuckItWeBall.jpg
>proceed to eyeball the image
>how many FUCKING VERTS ARE IN THE ARMPIT GOD DAMN I-
>decide to start a placeholder lowpoly model
>invertedly start making it higher poly out of a subliminal need for those glorious shoulder deformations
>"fuck this" goes for freely available models to use as placeholders cuz i've had the upbge installed through the launcher tool w/o even using the logic nodes yet.
>"hmm, this model is pretty great but it could use a touch up"
>wastes several more fucking hours trying to retopologize the shoulders, the arm muscles, the pelvic thigh creases,
>mfw.

Why does every anatomically correct topology model have to be paywalled. and why do i have to have the "everything HAS TO BE EXACT!!!" autism with ADHD on top of that...

File: Screenshot 2026-04-12 211941.png (1.6 MB)

1.6 MB PNG

my first object... (it's supposed to be a dog kennel thing after I fill it out more)

I just gained a lot of experience with parenting and unparenting, joining and unjoining, separating, and setting individual pivot origins based on cursor

File: box.png (1.9 MB)

1.9 MB PNG

>>1027535

File: 1730504185072190.jpg (29.3 KB)

29.3 KB JPG

>>1027550
Ok so I instead just re-downloaded the original model again and that exported fine as expected. I must have fucked something up in an edit ages ago that messed with a script or something.

what is the difference to blender 5.0 vs 4.1

>>1027564
https://developer.blender.org/docs/release_notes/

>>1027565
i know, i was mainly asking what is you guys' experience and opinions between the two

>>1027566
Took me a week to get used to it and redo some settings, overall it's better especially with Vulkan enabled, but I don't use EEVEE

>>1027566
well, 5.0 in general has been getting us closer and

closer to nodifying every aspect of blender
bundles for example basically tape several wires together, and output them as a single cable that can then be unwound where you need the property again, cleaning up a lot of crisscrossing noodles.

closures however, allow you to define a zone of node logic that then can be plugged into an evaluate closure node using one socket.

its a workflow perfect for swapping out behavior of a node group or the other zones rather than cutting it apart.

then there's the bone info node and the geometry attribute constraint, two steps towards node-based rigging, albeit still with some middlemen.
Those two features allow a myriad of things: bones that obey cloth physics without an addon, a unified workspace to modify similarly behaving bones by replicating constraint logic in geonodes so we dont have to click on each individual bone to tweak one setting in each of their constraints,

5.X is THE place to be for procedural workflows and its only getting stronger.

>>1027568
The bone info node is too limited. Bones need to be their own component with their own index. Similar to curves. But not exactly curves. But also capable of being converted into curves.

There should be an *armature* info node. It's should provide all the information that the bone info node provides. Pose, local pose, transform post, etc. But then it also includes a green output that displays bones when you connect to it. Then you can connect it to a split to instance nodes, if you want to make instances out of them. Or a bone to curve node, to make curves out of them.And in order to select a specific bone, you use the separate geometry node, and select the bone by index. Either by isolating the number, or by using a named attribute to pull the bone by name.

File: clouds vdb.png (1.8 MB)

1.8 MB PNG

Anons, I made clouds, but when I imported them, only the first one has a visible structure on Solid Display, plus a proper a proper hitbox where I can click and modify. The others I have to click them on the menu. Can this be fixed? why it happened?

>>1027593
I entirely agree with you. frankly I dont understand why armatures dont have a modifier tab all their own so that geonodes dont need all this rigameroll to interact with them.