Hi,

I am messing around with the new auto rig feature, and have figured most of it out. But the one thing I noticed is, I am importing a FBX file with multiple objects in. And need to skin them all, then attach the auto rig.

When selccting the subnet with all the objects, it is not happy.

If I use an OBJ, it works fine as one object.

The autorig seems to only want to associate with one geo object. Is there a way to skin and or attach multiple objects to the GeoFile?

Or put the objects in container to associate with?

I'm sure multiple geo files into a merge will work.

yes, merging all the geo you want to capture will work
I'll add an RFE for selecting multiple objects.
RFE #82567
added related HOM RFE # 82579

Oh yes, that worked for obj. I think I tried file merge, but not merge.

Also, the next question is. I am importing an FBX which brings the goemetry into Houdini and is not file based.
I noticed when I call a single FBX file from the autorig wizard, the path to the selected geometry object is opdef:../..?geo_skin.bgeo

Not referencing the object I want to select or change

But how would I call the the multiple geomtry or have multiple file node call to different geometry within houdini and not file level?

Is this possible?

The other thing I just thought about and am trying to figure out is that when assigning a material to the auto rigged geometry, it appears that the material is assigned to

geo_skin

which means that there is one meterial for multiple objects in this case. So I am assuming an single obj with multiple objects there is probably a way to use different channels in one material.

But for multiple obects, how would you assign multiple materials so that different objects have different materials.

for FBX or for multiple objects in Houdini the easiest way would be to make a new Geo Object and object merge all other objects into this new object.
once all the geometry is contained in one object you can assign point or prim groups etc.
then in the Autorigs you can point it to this new object.

since all the geometry is in one object, if you assign a material to that object it will affect everything, instead you can go into SOP level and use a material SOP to assign different materials to different groups.

see the simple example file.

Oh yes, I see now. That is very helpful.

Thank you Arctor

I am trying to figure out how to associate motion capture to the auto rig.
Is there a module for this, or perhaps anyone know of any documentation to do this for the new auto rig features?

there are currently no tools to easily support mocap data in the autorigs.

Oh man, that sucks. The Autorig worked really well. Especially with the skinning envolopes. I didn't even have to edit them.

Trying to export the rig in an FBX and bring it into another package didn't work all to well, although that could possibly be my settings on import/export

Hopefully there will be future functionality for this.

Looking forward to being able to do that in Houdini!

vertigo262
Trying to export the rig in an FBX and bring it into another package didn't work all to well, although that could possibly be my settings on import/export

Hopefully there will be future functionality for this.

we are looking into these workflows, keep an eye out and don't be shy about asking for features etc!

oooooooooooooooooooohhhhhhhhhhh!

You want my input? lol

ok, Mantra should be CPU/GPU renderer. I am using redshift, and I love it! but, things always work better when Native, and less bloatware.

Also, this is very important. The viewport should support soft shadows, VXGI real time global illumination, atmospherics, particles

And have the ability to, instead of heavy rendering, capture realtime frame grabs, like you would a photorealistic game engines.

In other words, You should set up your animation sequence, and render off the viewport display like you would a game engine because I am guessing with where technology is at, in about 6 years. People are going to be doing a lot less rendering due to the fact that Game Engine technology with GI capabilities is now starting to get close to render quality. Minus the amount of GI bounces it can take.

So instead of using a game engine to create animated sequences which people are starting to do. And I've seen some amazing realtime GI animations, You just grab Houdini, Animation Software, and render out your real time sequences provided the Viewport is capable of doing what game engines like Unity or Unreal and many more are doing.

But game engine interfaces, at least the ones I've messed around with, don't really have the ergonomics for animation. Not saying they can't do it, just don't usually flow right.

They tried this in Maya years ago, but the GPU technology wasn't ready. Now it is.

That's my 2 cents, because that is the future of CGI!

Anyways, that's just off the top of my head

https://www.sidefx.com/bugs/submit/ [sidefx.com]