Rigging and Modeling with Multiple Artists?

Artistic Support Animation and Rigging
, ,
1

Hey,

I’m sure this has been asked before but I haven’t found anything to help me out too well. I’m wondering if it’s at all possible to have a modeler and rigger working on the same asset at the same time. Or at least have the modeler be able to update or fix things in the asset and have it not completely remove the riggers work.

Currently I’ve tried linking in a model and creating a library override to be able to paint weights. If I then go back and alter the model even slightly when I go back into the rigging file the model is updated but all the vertex groups have been removed.

Is the only solution to work in a single file?

I’ve actually seen the Blender studios Asset Pipeline addon but that doesn’t seem to load for me. It looks like it installs fine but it doesn’t appear in Blender or the addon menu.

Thanks

2

When you go back to alter the model, is it the one with vertex groups or without vertex groups? Does the modeler add the vertex groups or does the rigger do it?

3

The rigger adds the vertex groups

4

Hello !

First this is a workflow that works in Maya, but in blender it’s more difficult, some people tried it and I’m not sure they have been successful.

A completely different approach to solving that is multiple people working on the same scene : https://gitlab.com/slumber/multi-user

To have discussed with pipelines TD on different maya based studios, some would argue that its required to build complex projects, some other say that it’s prone to error and should be avoided.

My take, is that especially with blender this isn’t something you should do because blender isn’t really thought around that. And therefore you’ll build a complex methodology just to circumvent this limitation, creating more trouble than it’s worth down the line.

In the end, if some modeling work needs to be done while the rigger is also working on the scene. Just let the modeler work on a copy of the scene, and then the rigger will append the changes.
It’s something that should be more the exception than the norm.

Sometimes you are forced to work rigging while modeling, and it’s something you should avoid. Because each new iteration of the model , proportions would change, you can’t make vertex groups unless the model is final and so on..
It’s basically a waste of time that should be done only if you can’t otherwise.

In a more classic project, you can just schedule the work in a way that people don’t step into each other toes, and when they do it can be handled with basic tools. Even the most robust technique will break if you change topology, and transferring weights is probably not going to be good enough.

As long as your team is small enough so the modeler can talk with the rigger, then this is likely to be much more efficient that any clever tool you can build !

Happy to discuss that further and share production experience as a learning opportunity !

5

As @sozap said that workflow is difficult with Blender because the program doesn’t have the needed functionality natively.

At a studio I worked at some time ago they developed an addon that instead of linking a file it would append the model to have it completely local, and it would automatically check everyday if the original file where it was appended from had any changes. If so, then the addon would append the latest model again, it would copy any vertex groups (with the data transfer modifier) from the previous model to the newly appended one then and delete the oldest one.
It was cool and practical because it automated a lot of repetitive tasks under the hood, but we were forced to constantly tweak and retouch weights anyway, because the data transfer from one model to the other would almost always mess something up, especially when big changes in topology were made.

So the rigging team was spending way more time than usual on each character/prop because the model was not “locked” before starting the rigging process.
By the end of that production our supervisor instructed us to just do a very basic version of the rig using the first version of the model as a reference, make sure that the main functionality was there, and only move to weight painting and shapekeys after the model was fully approved.

And that’s the process we use at my current job too, the rigging team receives a blockout of the character/prop to start building the metarig and testing functionality, but it only finishes the rig and weight paint when the modeling team delivers the final, approved version of the model.
The only 2 processes that are done at the same time are texturing and rigging, because they both use the same final model and there’s no risk of messing up the other person’s work.
Modeling and rigging at the same time just creates more problems than it solves, especially for productions where you have tight deadlines.

6

There are ways to do it but generally, I would avoid it. We’re by no means a major studio but we have stage locks at work where, once a model is marked finished and has been approved, we don’t make any more changes to the model (unless something is broken). Once a model is sent to riggers, unless it’s broken, no one, not even the boss, can ask for model changes anymore.

Now, that may not work for everyone. We have very tight deadlines- months, not years, for a major project start to finish.