With the imminent release of version 3.5, and soon after this, 4.0, I’d like to reopen the topic of using Z-spheres as the ideal basis and ultimate realization of the most efficient use of the computer for many kinds of rich animation.
From what I can see demonstrated regarding the latest functionality contained in Z-Spheres II, and their associated display and manipulation capabilities, I think we are a hair’s breadth away from the most interesting animation solution, streamlining the entire process for artists, that has ever been introduced to any CG audience.
After first being exposed to Z-Spheres, with the very first versions of ZBrush, I could see the obvious direction Mr. Alon was intending to go, ultimately, with this incredible sphere based technology. I can’t express the possibilities any better than have already been expressed by this user, way back in 2004:
ZSpheres - the future of animation!
Hi all,I just wanted to get a discussion started on the potential of ZSpheres (or something like it in animation). I think it could really change the face of character animation forever. The problem is that ZBrush is not an animation package. I’m not necessarily suggesting that ZBrush becomes an animation package - I think it’s already easily the most revolutionary modeler ever made and I want it to stay on top as a modeling package, but I think that other packages should start exploring alternative ZSphere-like ways of animation.
The traditional problem with animating complex characters is handling a huge number of vertices and making them deform properly. People are exploring methods to allow them to animate low res cages but still keep the high res details intact through pixel shaders, displacement, etc, but I think the true future lies in hierarchical subdivision surfaces where the artist can work with any level of detail he likes and have the full geometry there to manipulate without having to generate the remaining level of detail through one of the aforementioned techniques.
ZBrush is brilliant because, for one thing, it supports HSDS better than any other package I’ve seen with practically no difference in speed from a subdivided model of matching resolutions.
However, even if a package supported HSDS as well as ZBrush did, the workflow for building a character is still complicated by the traditional system of using bones to deform characters because of the nature in which bones deform geometry. The problem with bones as an underlying system of deforming geometry is that you still have to have specify which vertices are under a bone’s influence and to what extent. As a magnet-like tool for pulling points and rotating them, there’s only so much you can do with bones. The weighting of vertices ends up becoming a burden and, even with excellent influences on each point, you still end up rounding off/bulging areas because, alas, there’s only so much you can do with an underlying magnetic system.
Enter ZSpheres.
The amazing potential ZSpheres has to me over the traditional bone system is that it works as a hybrid between the low level control cage and the bones. It is a jointed system which you can manipulate like bones, but it is the geometry in itself! It’s has the ease of manipulating a segmented character and you aren’t so much deforming the geometry as manipulating the geometry directly. You don’t have to worry about the problems of deforming bulk because the bulk is part of the ZSpheres themselves! It’s like metaballs-meets-bones-meets-HSDS.
You can forget weight mapping under this system, and corrective morphing should become a much lesser need.
Anyway, I just wanted to get this idea thread started. Imagine something along the lines of ZSpheres in an animation with IK goals, rotational limits, etc.! As an FK system, I see no reason why they can’t have this.
The entire thread can be read here: http://www.zbrushcentral.com/zbc/showthread.php?t=22150
I’m not expecting any of these things to be released in 3.5, but I’m sure many old time users may be thinking along these lines, and I want to encourage as many users as possible to do the same, as well.
Sincerely,
Greg Smith