Building a model from smaller rigged models?

[williamgaylord]

Here is an interesting challenge: Make a complex rigged model out of smaller rigged models by importing the smaller models into a new model. Like a cut and paste operation, it could enable you to rapidly build complex abstract or natural looking models (like trees) to be animated in fascinating ways. Iterative operations like this could get you into trouble, but done right it could make for some amazing possibilities!

 

I don't know if it will work, especially importing models with constraints and relationships applied. I would think that all the parts, constraints, relationships, etc. would need to be renamed to be unique each time prior to being imported. (I've tried it without renaming and got a mess.)

 

Any ideas on how to make this work? Can it be done with AM in it's current incarnation? Perhaps a mix of using a text editor on the model file and the project workspace tools?

 

I would start with very simple abstract rigged models with simple rigging for a test. I'll try something myself this weekend. Feel free to contribute! Especially if you've already figured it out!

 

Bill Gaylord

[nyahkitty]

Sounds promising.

 

One strategy some employ is to build separate parts of a model and then bring them into the main model (arms and legs and such). This has usually been regarding just the model mesh.

 

I often like to take that one step further on high patch counts (5,000 - 30,000+) by creating a new model as a sort of temporary tray to hold part of the larger model in order to work on just that part in isolation. I'll copy/paste the part into the new modeling window. When I'm done, I archive the main model as a file version, delete the part that I'm replacing, copy the part from the new model window and paste it into the original, larger model. Then I just tidy up any redundancies in the Groups folder and Decals folder.

 

To get bones and perhaps even assigned constraint poses in there from a Temp tray, you are correct that you would need to use the Import Model function. I'm looking forward to seeing what kind of workflow you discover as well as what situations this would be useful.

 

Happy experimenting!

[williamgaylord]

I think I discovered why the iterative approach doesn't work. Only the spline and control points are automatically renamed. If you rename what you can in the model to save it as a unique model, then import it into a new model as I just did, the names will be unique, but the references in the constraints to splines or control points, etc. will not change, so they will conflict, causing a "circularity error".

 

Note the identical spline reference in both imported models in the attachment below.

 

Maybe there might be a clever way to avoid this be editing the relationship references in the file, but how you might track how the import process changes the numbering in the mesh would be the trick. If it just starts the numbering from the last one used, then you should be able to just add an integer offset to all the constraint references...or something like that...I don't know the syntax of AM files yet.

 

Anybody have any ideas on how to overcome this?

 

Bill Gaylord

[johnl3d]

just random thought can you assemble the pieces in the chor and save the model assembly from there some how ?

[pleavens]

Have you tried assembling them as action objects added to a base model? And if so, were there any issues you discovered?

 

Phil

[seven]

The programmers on Hash ought to know!

But I never see any - Bob Croucher - taking parts in the discussions here!

 

Because if this will function it would be very interesting...

Nature - the cells - are in some way constructed in this iterative way... or better dialectical way:

http://en.wikipedia.org/wiki/Dialectical_materialism

 

And doing this kind of animations - rigging - would produce some kind of virtual reality animations... making it possible for us to visualize some ideas... it can be nice... but I also see the dangers in it... but people... corporations are doing so many dangerous things these - last? - days - playing with the genes and nanotechnique - so why not play a little in A:M with smart iterative rigging?

[williamgaylord]

just random thought can you assemble the pieces in the chor and save the model assembly from there some how ?

 

I usually build a trunk with a few branches as the base model. I then build a branch model and then save it a few times to make a handful of branches I tweak to make them look different. Each of these has its growth control reduced to a percentage pose slider. I rename them as well. Then I bring the models into a choreography to assemble a tree using multiple copies each of the branch models. It assembles as the spline paths that guide the growth, each branch part growing at the proper time, which I adjust in the timeline.

 

Now, if there is a way to save the fully assembled tree from the chor and have it work as a single model, that would be almost as good as the iterative idea!

 

Have you tried assembling them as action objects added to a base model? And if so, were there any issues you discovered?

 

I have heard of action objects, but don't know how they work. My impression was that they were sort of like proxy models that can hold constraints, relationships, etc., so they can be applied to other models. Tell me more.

 

Thanks!

Bill Gaylord

 

 

 

 

 

 

Bill Gaylord