this is a great research project even when it will take a few years to catch up with zbrush on this level.
in terms of voxel sculpting i think extremely special care has to be taken to make sure the tools work well, thats my biggest gripe so far with 3d–coat. unlike Z-brush where all of the tools seem to work so incredibly well.
funny thing though is that its hard to specifically define what the difference is, i think it would be a good discussion to try and figure out good workflows and tool use when voxel sculpting.
what tools do other people think are necessary for voxel sculpting?
What about being able to convert envelopes to a voxel-mesh?
From what I understand with 3d-coat is that it only has the area that your brush is making contact with as voxels, the rest of the model is a generated mesh, which is how it accomplishes such highly detailed models using voxels but the cost is the quirks and messed up meshes; not to mention that it crashes if your brush size is too big. So that’s both good and bad I suppose, good in that it allows high levels of complexity but bad in that it’s quirky.
ToastBusters, have you worked with 3d-coat? I didn’t really experience this kind of stuff. Also, although there is a generated mesh surface, the full voxel structure is stored.
The full voxel is stored,but it must be tessellated.
Voxel has this disadvantage.
It’s a dicretized surface,against an analytical surface(in some way).
A polygon mesh is an arbitrary and linear(the edges are lines)surface,even at the lowest level if you displace some vertices you have smooth(not curved of course)result,from point A to point B.
When subdivided with a cubic subdivision scheme,it’s actually a curved surface at the limit.
In voxel you have stepped result,only reducing the step size give the illusion of smooth result,but for example if you model a plane not aligned with the world axis,you will never do it clean(a surface/mesh can do plane surface in a perfect way,and can approximate in a good way at the limit curved shapes,voxel will always be stepped,only at high res it can look good)
well, your post here came without much info, but just for those interested here’s a link:
http://www.voxelogic.com/index.php?option=com_content&view=article&id=1&Itemid=21
Saw this the other day and was floored. Such fast progress! Amazing, farsthary!
I once tried to sculpt a 100x100x100 cube of vertices, but the Sculpt Tool couldn’t move individual points (1D). The Proportional Edit tool worked, though. I guess the tricky part is the screen drawing, and skinning it into a surface mesh.
Can’t wait to test it!
(BTW blaize, How’d you post something with less than 10 characters?)
It’s a start for sure.
I too wonder if you’ll be able to make it into a mesh, for things like the modifiers, unless of course he creates a ‘voxel’ mode for most of the existing modifiers so we can use a string of displacement modifiers to create landscapes with overhangs and caves.
Trianglization is also going to be important if you want people from Max, Maya, and XSI to use Blender voxel sculpting because there’s no format that will take pure voxel data that I’m aware of.
There are some great papers about anasitropic quad dominant remeshing. This alone should make retopo more or less obsolete.
Cool, I like these advanced papers even if I don’t understand most of it and primarily look at the pictures 
Here’s one PDF I found:
http://www.geometry.caltech.edu/pubs/ACDLD03.pdf
Searching for “Direct Anisotropic Quad-Dominant Remeshing” gave this as a first hit:
This is not a gratis paper though since it costs $19. I would be willing to donate this paper to a dev, but would it be totally wrong of me if I redistribute the .pdf to more devs, like farsthary, gensher, broken etc.? :o That’s also assuming this paper will be useful to them…
Just an idea, maybe something like “Voxel Patches”, something similar to Zbrush HD sculpting?
You have your low res polygon, then activate the polygons that lie within a radius from the mouse cursor. Then the voxels are displayed just for the active polygons.
This could also expand into polygroups. Maybe, a polygroups in outline?
San, the first paper is what I meant.
edit: I think your second link ($19) point to the same research. The greedy BASTARDS.
Furthermore, the work that Daniel Genrich did before the smoke simulation looks a lot like what the paper is describing judging by the results. Daniel said it really needs BMESH though.
Farsthary just posted his tentative roadmap for the Bloxel project (Blender voxels :D). Check it out:
http://farsthary.wordpress.com/2009/10/05/bloxel-blender-voxel-system-roadmap/