Pete the Horse questions

Pedrojhiann

Ok, thank you, Andy, but I think I wasn't clear on what I was asking.

How then do you bake the simulation onto the meshes within the file itself?

Cheers!

andy_vs

I'm not quite sure what you mean by 'bake the simulation onto the meshes'

If you're simulating the muscles, you select the muscles, and export the selection to alembic. This runs the simulation and exports the alembic file to disk.

When you simulate the fascia, you import the alembic cache of the muscles, and make a blendshape from every baked muscle to the corresponding geometry in the fasciaSetup.

Pedrojhiann

Thanks, Andy,

I wasn't aware that when you export the Alembic Ziva automatically runs the simulation for you. That was the piece of information I was missing. I will try this tomorrow and tell you how it goes!

Cheers!

andy_vs

Pedrojhiann

Ah! Great. Happy to help

Pedrojhiann

Hi, again, Andy, I hope you had an awesome New Year!

So, after the Alembic Issue was solved everything worked smoothly so far. I just have a small question regarding the final skin blend: to which mesh from the puppetSetup file I blend the wrinkle Alembic with? Also, do you have any suggestions as to which deformer to use? Blendshape doesn't seem to work as they all have different polycount and polyorders and shrinkWrap doesn't seem like a viable alternative either since it is too costly and the results are not really good from the tests I have run...

Cheers!

andy_vs

Pedrojhiann

Hey so great to hear things are progressing well at your end.

Yeah you still use a blendshape, just wrap a copy of the skin to the output of your wrinkle alembic.

So you have the following in your scene:

The output animated skin (from the puppet)
The output from the fat/wrinkle solve(a copy of the skin is wrapped to this)
A blendshape between the animated skin and the wrapped skin.
You just paint the blendshape to white where you want to see the influence of each pass.

Pedrojhiann

yeah, but in the Pete the Horse Files every skin mesh in the puppetSetup.ma doesn't match the Skin CLoth mesh from the skinSetup.ma. Any suggestons?

MAybe replace the skin Cloth Mesh in the skinSetup.ma with the puppet skin mesh from the puppetSetup.ma and run the simulation again?

andy_vs

Perhaps I don't understand your problem.

From my perspective they don't need to match. Wrap a copy of the skin (in this case the puppet skin) to the output of the skinSetup.

Then they match, make a blendshape between the puppet skin and the wrapped skin.

andy_vs

Or did you mean they don't even match spatially?

Then I'll have to take another look at the files

Pedrojhiann

Hey, Andy,

No, it was just the topology that was different, but anyways I got it working now, thank you!

The thing was that the Wrap deformer was just plain crashing on me and ShrinkWrap was just plain awful with the results, but I tweaked the Wrap a little and, presto!!! Now it works. Then it was just a matter of making a blendshape between the original puppet animation skin and the warped mesh. Thanks to your last comments I understood that the warping and the Blendshapes were two consecutive steps and not alternatives to each other.

Cheers!

andy_vs

Pedrojhiann

Great stuff Pedro thanks for the update!

Pedrojhiann

Hey, Andy,

I would like to show you a comparison between the Ziva Skin Simulation and the projected mesh. All meshes used were the ones provided along with the Pete files. As you can see in the video there are quite a few details that are lost in translation. On the other hand, though, the wrinkling seems too extreme in the Ziva Sim, and in this regard the more subtle effect in the projected mesh seems more naturalistic.

Is this difference in results expected? And do you have any suggestions as to how to close the gap in terms of detailing in the final result?

Cheers!

Pedrojhiann

well, can`t seem to be able to upload the video.

Pedrojhiann

Ended up sending you an email, Andy, if That's OK...

andy_vs

Pedrojhiann

What most people do is make a Vimeo/YouTube account and just post the link to the uploaded video. The platform automatically previews the video for readers.

Pedrojhiann

Yeah, Andy, I don`t know if my company would approve of me doing that on Vimeo or Youtube, though... I spoke to my supervisor and doing it within the forum would be OK at this point, but things would be trickier on more public outlets. But thank you for replying my email.

Now I have another question. Is it Ok for Ziva to use Negative frames as start frames?

And while on this subject, the proper use to apply Animation Data from elsewhere (and that starts out on a different pose than the Bind Pose) on Ziva would be to have all joints be initially on the same positions as the Ziva Meshes and gradually blend the positions over N frames, right?

Cheers!

andy_vs

Pedrojhiann

Hey Pedro,

No problem, just thought I'd mention it!
I haven't seen any issues with users running sims on frames lower than zero.

Yes you're right. That gradual blend is generally called a run up. Ideally this would be an automated process, when the animator publishes animation, the run up is automatically added to the start of the published animation.

When doing the run up, you can move the character from origin to world space over 1 frame, so long as you move the zSolver locator itself at the same time. I would recommend doing this on 1 substep. (substeps are keyable attributes)
Once the character is in position (in world space) you can solve into the animation start pose over 10 or 15 frames.

Hope that makes sense

Pedrojhiann

Hi, Again, Andy,

is there a way to set up distributed simulations with Ziva on, say, a Render Farm? Even with a single license?

Cheers

andy_vs

Best to use batch licenses for farm distribution.

You could run a 1 node renderfarm with one standard license but you wouldn’t be able to author while baking, which would probably defeat the purpose 🙂

« Previous Page