also btw what i write about overthere is all computed in the shader (which means per pixel) so it is more smooth, but the attributes themselves like the spline position, curve tangent and curve factor are calculated in GN and main calculation with the position [object coordinates] (which is actually per pixel) is in the shader, so it is pretty good, anyway.
Is there a way to merge curves in Geo Nodes? I have this text made by āStrings to Curveā node that creates a hole at the area of overlapping when the curves are filled, and that is not what I want.
can you seperate the two curves, fill each seperately (which creates meshes), then if you need them as a curve, boolean those two, and mesh ā curve?
4 Likes
with arctan2 you shouldnāt have to normalize anything (except the normal)
Hello, about to buy them.
Are there any build up nodes? Like for example when we have an object with many different objects in a collection that we can animate them in, in a cool way. I build that now several times and should put that into my asset folder with some improvements but i probably have not chosen the best way.
Hey, by build up nodes you mean similar effect like this?
If so I have falloffs nodes to help achieve similar results. You can find blend like with falloffs nodes to try here:
8 Likes
Thats exactly the type of animations i mean. Possible to do in an endless way and very customizable.
Thanks!
Hello,
First of all Iād like to say that this collection of GN is the best Iāve ever seen, it opens a huge amount of creative possibilities ā thanks Higgsas for the amazing work!
My question is: is it possible to adapt the Circle Packing Solver to pack spheres with different radius inside 3D objects like spheres or cubes?
Iāve seen a few tutorials where the packing is not āfullā, I mean, thereās always empty spaces among the spheres.
What Iām looking for is a way to do the same shown in the video below (not necessarily animated):
Hi. Iāve been searching for similar think like push apart effector in c4d and make this setup
push_apart.blend (1.5 MB)
You need blender wtih repeat zone. There is 2 tweaks - iterations (10-30 is better) and distance multiplier. It calculates distance between points in 3d, but not objects boundaries. So if there is a spheres with random radius, there will be gaps between them.
2 Likes
Hi, thanks.
Yes, you just need distribute points inside the volume and run circle pack solver.
circl_pack_3d.blend (166.2 KB)
5 Likes
would this (mebbe with minor adjustments) work for non spherical objects too? for example, random rectangular objects? or even differently sized monkey-heads 
? or should i stick with rigid body sims for doing that kinda stuff?
after you have the ābasisā there is a way to sim now also @KDLynch
we need a bunch of nodes though.
1.lists
2.gather contacts per physics shape
3.apply impulse
we can then do a rigid body sim right in blender
if you need it right now - there is the upbge fork
(if you need it animated I have it covered also)
record.blend (850.9 KB) - this records gameplay to f-curve animations when you quit
2 Likes
Amazing ā thank you very much! Hope Blender 4.0 brings you more inspiration and tools for new node groups!
1 Like
Hi, is it possible to create a mesh deform modifier in GN ?
Hey, you can try this node group itās pretty similar to mesh deform
1 Like
Useful tip: in recent versions there is a new feature in mesh properties - āAdd Rest Positionā. If you are using input geometry you have access to ārest_positionā named vector attribute that freezes incoming geometry
(discovered this in Bcon2023 video)
2 Likes
Have you consider making intestines. they follow the same principles as the reactively diffused brain with only two key different.
- There must be a start and end point that doesnāt move
- it must a single continuant line
I guess there the whole issue that it must also be connected to a membrane āskirtā that goes to spines that also effects the geometry of the intestine too.
Hi higgsas
I try to use your cool curve deform node to make a fence with posts. Can it be modified the way the posts would stay vertical when deforming curve goes up and down ? As if curve normals would stay vertical but still getting some tilt around curve tangents ?
Hi, is not really possible to create just a single continuant line that has endpoints fixed, there is a different F and K values to use that would give more like continuant line but it will eventually split and you cant really make reaction diffusion work in 3d with current nodes.
But for the similar effect you can try differential growth solver, I donāt have it for now, but you can check Alex Martinelli differential growth addon:
1 Like