the quad mesh looks great !
time to take look at zbrush - simply in regards to the mesher …
the quad mesh looks great !
the problem i most often run into regarding the mesher is it creates long skinny triangles when i wish it could just do quads… (those triangles can have weird results in my renderer/textures)
for example, the unjoined surface on the right is how i wish it would mesh that section even when it’s a polysurface (left)…
i imagine it’s a bit of a nightmare to try to make a meshing algorithms that works right* for all/most occasions.
*with ‘right’ being a particular user’s expectations.
(that said, i can get more usable results than this with different settings… not as clean/low-poly as the version on the right but still, better)
A couple of months ago I played around with OpenFlipper, a mesh modeling program from academia, mainly used to do research around meshing algorithms. It can read and write various mesh formats and supports re-meshing. That is, you feed it a mesh and it will re-mesh it, to give a finer or coarser mesh, or with better shaped triangles (similar to Delaunay).
It is maybe useful to someone. To be honest it’s nowhere near Rhino in terms of usability, looks and stability, but it gives nice meshes.
As you’ll know this software is actually by the group which created several papers on quad-remeshing. But these algorythms aren’t included in OpenFlipper… Purely triangle based meshes on roundish shapes may have poor shading properties and were a nightmare for advanced concepts of texture mapping, so this route is limited.
Yes I know their papers. I’m not familiar with texture mapping and shading properties of triangle meshes, so thanks for clearing that up. If I’m not mistaken, their quad remeshing algorithms are also the basis of Evolute T.MAP, which is currently in beta - see Release of EvoluteTools T.MAP *beta*
The beta version has texture mapping, but no meshes.
Rhino could recognize untrimmed surfaces and create quads like like the UV directions. Sounds very simple.
I am working on a quad mesher for Rhino and should be ready announce something for Rhino WIP users next week. The quad mesher completely remeshes a surface or a mesh, and can have quads flow across patches ignoring patch boundaries.
Keep in mind that it is still just a prototype and not ready for release.
I tried remeshing the model posted here but there were too many issues with it e.g. several non-manifold faces. I need to work on making the quad mesher more tolerant of these things in the future.
Here are some screen shots of other models:
great news, strongly looking forward to it …
Wow, that’s what I call a surprise @DavidEranen!
Very nice how you got rid of the singularities on the helmet and created edgeloops around the holes.
Also the letters look great!
Are you also planning reprojection (via pull) so that one can create such a low res cage, which gets used for unwrapping + higher subD-levels which (as reprojected) very closely aproximate the Nurbs source and have inherited the UV-map?
And: As this can’t be the realtime rendermesh…Are there any thoughts in terms of transfering / linking this mesh to an updated Nurbs source (for the case that one cuts a third hole into the helmet, so that one hadn’t to start over from scratch).
That’s really great news!!
In case you need more test data, I can provide some challanges!
hifred, I’m glad you like it!
In regards to the different functionalities you’re wondering about: I can definitely see these features being implemented, but for the moment I need to focus on 1) making it work very well on many kinds of complex surfaces and 2) improving the algorithm speed, which can get really slow with larger models. There is still a lot of work to do on these fronts.
I will accept all test data I can get my hands on! For now I’d prefer smaller models over large ones, because it makes more sense to iron out issues with smaller ones first.
Wow, impressive!!! The mesh of the helmet is independent from the NURBS UV and looks perfect.
Yes, this is impressive, indeed. This would be an awesome tool to have within Rhino!
That’s smashing news, thanks!
This is fantastic!!
ReduceMesh can be a solution for you.
Here I extracted the original rendermesh and reduced it from 57701 to 4000 polygons:
This is so impressive for an early prototype.
This will be useful in so many ways!
I can’t wait to get my hands on it.
this is so amazing! i cannot wait! great!
Don’t get yourselves all worked up just yet - there’s a lot of problems to solve with it, and a lot of cases it doesn’t handle. It’s still going to take a fair amount of development.