In max, maya, softimage, unreal, unity…pretty much any 3d software you get the option to select some objects and treat them as one, and instance those groups. Except blender. And no, collections or parenting are no substitutes. But this is off topic.
1 Like
Anyway I feel I am dragging this thread down a little bit which is actually not what I want to do.
The positive thing is that, unless I am mistaken, this is the first time we are actually getting “real” animation nodes in the next master. Lets see what the future brings! 
I am certain the developers know what they are doing and while I can’t say I always understand the decisions being made I am certain they have our best interest at heart.
Love you devs! Sorry about being a little moody. 
4 Likes
Not sure what you mean, I do this all the time… but we are indeed off-topic.
So, what would you like to see in geometry nodes that isn’t part of the exposed plan ?
If you have a way to scatter things on a surface, you can scatter leaves on surface of a tree.
Tree generation is more about branching.
But if you have a way to scatter branches the same way, you scatter leaves. You just need to remesh to obtain a tree.
I don’t doubt that would be something that would rapidly become possible with addition of few additional steps.
Since a long time, we have seen addons generating trees from Lsystem, curves, hair particles.
I have no doubt that some studios or node enthusiasts would be interested to put some efforts to create nodes for tree and plants generation.
Well, currently, Displace modifier is using blender internal textures.
I think that target corresponds to the ability to have a displacement matching EEVEE/Cycles textures.
Let me clarify what I wanted to say. I think that people are not using particles for abstract cases.
Old system is not great for fireworks. But it is used for hair, fluids, boïds, scattering…
Demos of simulation nodes were fancy but not very practical.
Dynamics is particularly not easy to apprehend when you are used to old system.
Developers discussions were a lot about way to present new functionalities to be understandable by artists. And that is more in UIconcepts of nodes/sockets/nodegroups/node editor views that abstraction took place.
I think there were 3 or 4 proposals for simulation nodes design. And the way to use them is not tangible, yet for most of users.
So, that is a good thing to split it into several practical examples : set dressing, modeling, dynamics.
What does it matter what I’d like? I’m not here to make feature requests. All I hope for is that this will be an ongoing project and not another desiccated corpse of a feature.
If collections are not a substitute, yet ; it is because of absence of node system.
Collections are currently grouping for rendering.
They replaced groups in physics settings.
You can consider using a collection instance as an object and the fact to enabling/disabling corresponding collection like switching between edit/object mode of that object.
In 2.91, collections are becoming grouping for boolean modifier.
With geometry nodes, we will probably have a collection node that would amplify that idea to the whole procedural modeling and dynamics abilities of those nodes.
So,we should end-up with collections covering all grouping uses.
1 Like
Perhaps this has changed since a couple of years ago but the groups in Max are not usable besides for the simplest of tasks because they prone to make Max crash.
Got something animated inside the group? High chance of crashing.
Got some non trivial hierarchy inside the group? High chance of crashing.
Grouping in Max is (or was) so bad that a lot of people never used it and went the “link to empty” route.
1 Like
I am only inviting you to give some positivity by sharing what you think this project ought to look like, instead of bashing it into the ground. You could start by listing what you think their plan lacks, and how you would solve it. Of course you don’t have to do any of that, but then please don’t be so dismissive with me : I am a good man.
2 Likes
Hopefully that is not too far away. The current version of animation nodes that is out there already has an L-system node.
If I understand the blog and stream correctly if a volunteer would write an L-system node for the current infrastructure (don’t know the proper term) and submit it, it could be part of the next master release essentially allowing you to do so.
If they meet their goals which includes “flower scattering” I’d assume the node set would be complete enough to build setups that would scatter leaves on the branches too.
That really depends on whether they go low or high level with the nodes through. One of the things I wonder about is whether this focus on the next movie is going to lead to highly specialized nodes or actually proper low level nodes that expose every component and in turn allow you to build almost anything.
4 Likes
An Lsystem is one thing ! but then you’ve got to make it into a mesh with thickness and UVs so that bark can flow in the correct direction. You also want to have it sway in the wind by transforming the Lsystem nodes hierarchically, and you want scattered leaves to have some simple cloth dynamics. Finally the user should be able to pack it all into a tree asset and instance it over their scene, changing exposed parameters to create variations. That would be my ideal scenario.
1 Like
Animation Nodes has nothing directly to do with the current Particle Nodes and Geometry Nodes. It is a (mainly) python addon and is not integrated into Blender the same way as Particle Nodes or Geometry Nodes.
I am fully aware. 
Hence why I expressed my excitement a little further up about getting the first implementation of “real/native” animation nodes in the next master. In my mind geometry nodes (and particle nodes) are just a subset of the whole system. I guess I should have called the native version “everything nodes” to differentiate from “animation nodes” to avoid any confusion.
With that said there are obviously some lessons learned from the original animation nodes. Both from a usability perspective as well as what users managed to create with them. When you see a node such as the L-system node for animation nodes its a no brainer we can expect an equivalent for “everything nodes”. i.e. a native version. That is why I referenced it. 
1 Like
People expecting Blender to become a Houdini are going to be disappointed. I think that Blender should evolve into a better Blender, not becomig what other software already is. If you needs or wants a Houdini, there is already a Houdini out there. Probably what we can expect is a solid new foundation that opens two paths: an ever evolving system of nodes by the team based on their resources, or an ever evolving system of nodes by the community (lots of add-ons, shared node trees/groups, etc).
Also, it is too soon to judge, but I’m fairly optimistic. 
8 Likes
i feel like Geometric Nodes is realy more like a Nodebased modifier, than an actual “everything nodes” system. I mean Animation Nodes was more general, as it felt more like a Scene level nodesystem than a Mesh level one, like geometry nodes.
While this is really great, and the linking system inside the modifier looks promising. I just wonder how this helps with set dressing, beyond pebbles or leaves. It somehow feels weird to me having all that Set dressing Geometry inside a Modifier or an Object.
So i hope there is going to be also a scene level system in the future, where we can manipulate Scene Level Objects ( like for example the geometric nodes properties 
)
It is not meant to be Everything Nodes. It is meant to be the part of Everything Nodes concerned with geometry.
6 Likes
Don’t take my word for it, but I suspect there are going to be scene assembly nodes at some point in the future. Tangent has their own version working already as user&developer @bsavery showed in the USD thread.
3 Likes
Interesting if geo nodes would be compatible to images. And I NEED blur node.
I’m not bashing it into the ground… am I? I thought I was fairly restrained. Guess it didn’t come out that way, lol.
Anyway, there’s nothing explicitly wrong with this project, other than that it doesn’t seem like a good workout for the new node system. If they’re doing a movie project in support of animation nodes, perhaps it should be postponed until the node project is in a better state. You don’t really need to make a whole movie to test pebble scattering adequately. A smaller test scene like Epic sometimes make for UE would have sufficed for that. And once the nodes are in a better place, then launch a movie with the plan to generate as much content procedurally as possible. Trees, terrain, rocks, rivers, architecture, cloth etc.
As for specific features I’d like to see, I have no idea, I’m not familiar with the system yet, so I don’t know its shortcomings.
2 Likes
Geometry Nodevember!
1 Like
Just to note, not every feature produced in conjunction with a open movie withers on the vine. Motion Tracking for instance received quite a bit of work after Tears of Steel was released.
In addition, it used to be that the BF had to do fundraisers to even do these movies, so the sharp decline in resources after the movies were finished meant it was not possible to continue the work on everything they developed at once. Now of course one might cite Sintel as an example of new features that failed, but the failure of the render25 branch may be why we have Cycles now.
3 Likes