Freeform (Procedural) Animation: Modular Rigging -- A new approach

Geometry nodes are great imo. They get you probably 65% the functionality of Houdini for game assets. With the proper foundational framework, they could definitely be replicated in Unity and, if done well, they could push that number a bit higher.

There is one asset called Archimatix that goes pretty far in something like this, but it isn't as much 3d geometry as it is 2d sweeps of geometry over lines/plans into a semi-3d shape (i.e. lathe, but along a procedural spline/floorplan) meant for 'architecture' such as moulding, etc., (which is only one element of modeling) -- but overall, it is pretty flexible.
The biggest thing it is missing is 3d boolean operations. If you can sort that, texture projection, and handle voxels in a way that is flexible and convertible to mesh representations, you'd have a start to a pretty badass tool.
Archimatix is useful for simple geometry though (and some quality-of-life utilities for modeling static models). The texturing side is very iffy with this approach though. However, the biggest issue is Archimatix is rarely updated, and works poorly sometimes (as it tends to have lots of bugs, since the code is a mess on the backend, and it uses a lot of third-party libraries). Usability is also iffy, as learning it isn't straightforward. However, Houdini isn't much better (out of the box) in this regard -- It's just more powerful and flexible (once you learn what works -- and what doesn't).

Just my thoughts.

 

I'm trying to learn it. But also don't really know where I can apply it currently. I'm learning Blender. I've decided to not open Unity at all this month. Do some asset creation.

I figured I'd learn the rigging system. And, well, I think it is the missing knowledge that a person needs to understand animation rigging. If you have a "Unity only" mentality, which is what I had, I think you miss essential background knowledge of where these systems are derived from, and then you wonder why the deep explanation of stuff isn't provided, on how these systems interact.

The reason is that Unity is pulling ideas from different areas of 3D technology. It doesn't create any tech that is fundementally different from what anyone else is doing, in fact they're all doing the same thing. And the knowledge is shared among them.

This is all good research on my part. I've looked at Unreal, Blender, and now I know that they're all the same, kind of.

I really got to finish my chrachter in Blender. Pull it in to Unity, and get to designing the game. And then use animation rigging again, bringing it back to the roots of why we are in this forum.

Bloody systems. So hard to learn.

 

Let me put a bump here.
I have been slingshot ping ponging with rigging for the last 2 years.
I use Unity, Blender, Wings3D. I have imported numerous files into each from all over this planet.
I now have a mess.
What was the starting file?
What was the workflow chain?
Each model is different in its initial workflow.
There are no standards. I know this is the ugly statement.
Fully rigged is not a valid term. Is the human character rigged all the way to the finger tip targets? Or does one have to add this? And do the tools in the workflow really enable this to be done? If you got the model from somewhere else and you have to modify it then it becomes yours. Period. I hope you dont have to back track after this.
Are the normals facing correctly?
Are the bones rotated along the Z correctly?
Are the bones build along the Z axis?
These are things to check for but the nightmare still befalls you.
The starting point is 'In your search: at what date was the 'ANSWER' posted? One needs to pay close attention to the date!'

 

Wow. I had no idea anyone was still using Wings3D... ...and their website is still updated / has that same little dragon picture from the 90's too! D:

I can't help but think back to my previous statement here, lol -- Judging from the website, the only thing that has changed with Wings3D is the copyright date.

But in all seriousness though --
I agree information on rigging is very dated. There are very few people innovating in this area these days. There are a few though. Sadly most of these innovations are either pretty obscure -- or the tools are simply "in-house" and not shared with the public.

True -- but there are "standard approaches" to rigging various kinds of models.

Which, in general -- suck.

Reason being is that actual workflows for these tend to be software specific. Nowadays though, we're finally starting to get crossover in the workflow features department. Outside of bones and vertex weights, this hasn't yet been the case across most software.

Checking out Akeytsu, Houdini's KineFX, and Animation Rigging are about the best you'll find in terms of rigging for Unity. Blender is doable, with the right tweaks to your workflow (i.e. set snapping for your bone creation tool to "volume" for bones to be created in the center of the mesh volume, such as in the arms/legs, for quickly building limbs). Sadly, weight painting can still be a chore in Blender. Autorigging (like geodesic voxel binding or something like Mixamo's approach) is the way forward on the automatic weight-painting front -- Sadly, Mixamo is limited to humanoid models, unlike others with a more manual approach. With the right setup though, Houdini's (new) rigging system is the best of both worlds. It is very new though, so it might be more straightforward with other software right now, if you're not already familiar with Houdini.

Anyway, I hope this is a good primer to get people looking in the right places.

 

I had to see your response to realize I had forgot a key item to the 'Posting Date'.
This is in league with when the dev starts this journey.
The dev start date, the dev rigging dates of training, the finding of tools or packages, the date of the dev's idea.
Keep these in time line order and that promotes one sequence. Mix that order up in any fashion and you will make choices that WILL paint you into a corner or leave you without the correct answers or getting stuck in a proprietary path with only an ugly way out. The corner, without, or stuck leads to multiple projects with different, unfinished or broken non solutions.
And the coup de grace is the tutorials have not kept up with the releases.

This is not taught in schools. It is hinted at in project management training but doesn't get the attention it deserved because companies are too quick to hire those that are trained at that moment in the project's choice of tools, business model or project model. This leads to employees leaving, fired/replaced, investors getting burned or projects changing/canceled. Who comes to the rescue? The next Tom, Dick or Harry promising a fix. That is migrations(run like hell) or rewrite(start over but the timeline is already chewed up in the investor's/budget's view).
So there you have it. Choose wisely reader and know what you are embarking on before you sign the dotted line.
My last adventure was a Unity crash of a major project, learning Amplify Shader, Shader graph, Surforge, Blender/Wings, Unity, Autodesk fbx manipulations, Archimatix, Rigging through the whole mess, importing what ever I could get my hands on, stuck for one year in the Korean Game Review Board for a submission to Google Playstore, and 5 projects still in the que. Luckily I am a lone wolf so I don't have to answer to a board, people or pine. But the angst of seeing the dream clearly and having to sludge through reality still exists no matter what your idea, scheme, training or ego tell you who you think you are.
S)omething H)appens I)n T)ime, B)arking I)ntensely T)o C)reate H)avoc, C)olon R)estoration A)fter P)erstalsis, W)hen H)elp I)s N)ot E)nough, S)topping U)nder I)ntense C)haos I)n D)ire E)emergency.
Don't bang your head against the wall or beat up on others. Back off, take a walk or a shower or pay attention to someone else to pass the high tide.
Getting distracted in the midst of the quell will give the better answers every time.
All this after $1000 in assets.

 

Good news! -- Just an FYI:

Akeytsu is going to become free in the very near future. The company has already been sold. As it stands now, Akeytsu includes features such as automatic geodesic voxel binding to rigs (for automatic rigging -- a Maya feature) and a great weight painting setup, with good visualizations and easy mirroring of weights.
Houdini's KineFX is much better for rigging/weight painting more complex models than just humanoids, but it is a little bit strange in the UX (since you use nodes to reference the skeleton/rig/etc rather than an actual visual interface -- until you select the actual nodes you need, and then you have the visual interface). Akeytsu, however, is free, and is really good at actual (manual) animation workflows and automatic rigging for humanoids. You can also automatically place bones in a mesh's volume from any orthographic camera view. So more complex models are still possible in Akeytsu as well.

This was mainly just a PSA to anyone who wants a (very good!) free animation program that is NOT blender.