I have this pipeline that I am really trying to utilize that involves creating characters in their most simple shapes first in 3dsmax (constructing out of spheres and boxes and adjusting until proportions are spot on). After this I take the object into zbrush to create a unified skin for use in detail sculpting however this is (now suddenly) where I hit the wall.
At times the unified skin will be able to capture surfaces accurately and keep things like fingers apart and capture the curves created where two objects meet but once the figure has a certain level of detail or number of shapes it all seems to go wrong. Fingers begin melting together or suddenly multiple thumbs seem to be created behind the thumbs in the front view and facial features just don't appear.
I have checked and I don't have any symmetry on that could cause the multi thumb error and I have cranked the resolution all the way to 1024 and still can't get the volumes represented that I want and project all just stretches everything out all wrong and creates way more surface errors.
Is there any way to increase the resolution beyond 1024 in zbrush 3.5 or any other means of automatically generating a mesh surface that only interacts with the outward facing topology like a unified skin in either zbrush or 3dsmax 2010?
If I can get this figured out it will make character creation so much faster and I can focus on the "character" of a character in the sculpting phase and not worry about whether a pinky has become distorted or a forearms slightly longer than it should be (felt I needed to explain why I want to use this work flow.
Replies
If you can get yourself to a more update version of zbrush (its free either way), dynamesh also helps speed up the remesh+reprojecting process.
On top of this, these strange planar surfaces become generated as though whatever method zbrush uses to create a unified skin simply couldn't represent the curves as visible so it instead created a box like edge.
I feel like there is some way create a continuous mesh from multiple objects based on only their visible surfaces but I just can't seem to get it.
The closest result I have gotten in principle involved me in effect faking a 3d scan of my model in 3dsmax using the david3d scanner software ( I literally set up a virtual space with the reference boards and created an Omni light and squashed it into a laser and rendered a video of the light slowly going over the surface). This produced the proper concept of " I only want what I can see as geometry" however; I would need to render several hours of video from each angle in order to get the geometry represented in a passable or sculpt-able state.
Any suggestions??
Just to make my issue clear again I'm looking for a way of making only the visible surfaces of a model comprised of dozens of elements into one continuous mesh.
Unified skin will try and keep the faces consistent across the entire mesh. If you need more definition along the head, you could try to subdivide it on its own afterwards.
That could be a setting as well (I can't verify without a screenshot illustrating it). But I know you can turn the smoothing off on the skin, to create a lego-y, voxel type look. Ryan Kingslien had a pretty good video that demonstrats whats going on in the background
For what it's worth, remesh needs to be used along with project all, and may need to be subdivided prior to the projection.
Also, again dynamesh.
Not the greatest examples, but: