Poseable ZSphere Facial Features

[Bruce Gregory]

The further quest for creating a posable ZSphere face has resulted in this. It seems there is nowhere to go from here in generating a comprehensible smooth mesh (adaptive) which blends seamlessly with these posable features. Because adaptive meshes are always recalculating themselves based on relative sizing and positioning, as well as connective density, these feature will not retain their mesh form as represented here after adding more "membrane" material in the form of more ZSpheres. It seems that for something like this to work, an additional option would need to be added to the ZSphere modifiers pallete that would allow certain areas of the adaptive mesh to melt together (ala metaball technology) or something new and original. I think this would be the ultimate ZSphere development, and provide the means for characters to be modeled, textured, placed in a scene and animated, simply, I might add, all within our native ZBrush environment. Anyone else interested in seeing this happen? Are my conclusions valid?

Thank you very much,

Bruce Gregory

http://www2.zbrushcentral.com/zbc_uploads/user_image-1031167428fbx.jpg

[tishinina]

I think that is a great idea. :tu: I hope it happens soon. :D

[Jaycephus]

Uh oh, Robo's gonna get ahold of this technology, and then... ?

[BluePage]

I am guessing, haven't tried it yet, but it might be possible to use adaptive skinning in this manner

1. Create start pose making sure of the skin
2. Make a End Pose making sure of the skin

then using zscript morph between the 2 poses using interpolation of each zsphere. Say for instance I want 5 morphing changes, each frame would be calculated on a per zshere basis. Each would move in time until the end. At each stage it would produce a skin and perhaps allow you to pause and adjust zspheres that don't come out right.

As for real-time animation in zbrush I believe you are right, not going to happen with 1.5. But how bout version 2? or 3?

[aurick]

Adaptive skins can be reposed and reskinned with the same number of vertices and the same UV coordinates provided you don't add anymore spheres.

The trick is to start with a model that has enough spheres to give you the movable points that you need for your animation. In this case, that probably means a few more around the eyes and the mouth.

Once you have enough spheres, you can skin the model, texture it, and then go back to move spheres around to your heart's content. Any adaptive mesh made with the same settings as the first one will still be able to use the same texture. This means that you should also be able to use all of them in an animation program as morph targets. Or you can create a frame by frame movie in ZBrush and addemble it as an animate gif or a Flash presentation.

[Bruce Gregory]

Matthew:

Yes it should be possible, and here "should" is the key word. I have spent a sizeable number of hours working with this concept. My experience, (and I may be shortsighted here), has shown me that to keep the detail as illustrated in my above post, you need to begin at the nose and work outwards. No matter what arrangement I have tried, (and I'm sure there are some that I haven't), progressing from this point yield creases and such which do not lend attractively to a good animatable mesh. And, as you know, as long as the key spheres stay where they are, the mesh is relatively stable. But I estimate that in some more extreme poses, the adaptive mesh will change substantially and not look like what you initially intended. I would sure like to be proven wrong on this, but so far, no takers.

Thank you very much,

Bruce Gregory