It's not without its initial flaws by any means, and I admit I gave up early because I wasn't even sure if a seamless result could even be achieved when using anything other than a depthless plane (or meshes that aren't meant to connect up). But it was producing results that were close enough that I could import the result into Softimage and weld boundary edges without a problem. I'll probably go back to it and try some other adjustments now to see if I can avoid the extra step
I should probably note that it also helps to use some kind of aligner mesh first so you can make sure the faces are all pointing in the same direction (isolate them and use tool: geometry: spin edge if they are not)
Also, I'm guessing that the pattern has to tile seamlessly within a square. So if you're continuing to test your own or other patterns, you might need to make sure that the edges of the rings touch the boundaries of the square that the pattern creates.
I found my first problem was that the constructed ring had to take up the same area, or just a tiny bit bigger, as the origin ring from which I'd duplicated parts of the construct. Otherwise it would overlap on tiling. Probably it would be better to make inside maya/max/xyz, seems like some kind of zbrush xform issue?
But then after that I get this every 3rd poly flip/rotate on the mesh, but not on a plane. I had deleted polys from the extracted mesh, so possibly this screws something with the verts/polys in ZBrush.
It's a necessary evil (I have a third similar pattern, and it too lines up as long as the faces do). Align edge might take care of some of it automatically, but for most meshes you'll have to get your hands a bit dirty if you want good results (not to mention time spent making sure the faces are nice and evenly spaced to prevent warping, etc). At least you're not linking every single ring by hand, and can select large swatches to spin at once.
Spliting the mesh up could help (something I usually do with micromesh geometry out of habit, since it can skyrocket the vertex count faster than a subdivision). Also consider the portion of the model you're sculpting; sometimes noisemaker might produce a result that serves well enough, or an insert mesh.
Ok, Spin edge ;P. Alright thanks for all the input and coming back, appreciate it greatly!
And for the sake of interest, got it to this stage, but the rings are too small, a step down and they're too big, hah.
Replies
http://cryrid.com/3d/zbrush/multimesh/
It's not without its initial flaws by any means, and I admit I gave up early because I wasn't even sure if a seamless result could even be achieved when using anything other than a depthless plane (or meshes that aren't meant to connect up). But it was producing results that were close enough that I could import the result into Softimage and weld boundary edges without a problem. I'll probably go back to it and try some other adjustments now to see if I can avoid the extra step
I should probably note that it also helps to use some kind of aligner mesh first so you can make sure the faces are all pointing in the same direction (isolate them and use tool: geometry: spin edge if they are not)
Also, I'm guessing that the pattern has to tile seamlessly within a square. So if you're continuing to test your own or other patterns, you might need to make sure that the edges of the rings touch the boundaries of the square that the pattern creates.
I found my first problem was that the constructed ring had to take up the same area, or just a tiny bit bigger, as the origin ring from which I'd duplicated parts of the construct. Otherwise it would overlap on tiling. Probably it would be better to make inside maya/max/xyz, seems like some kind of zbrush xform issue?
But then after that I get this every 3rd poly flip/rotate on the mesh, but not on a plane. I had deleted polys from the extracted mesh, so possibly this screws something with the verts/polys in ZBrush.
Edit: It's Qremesher that screws it. It's fine to delete polys from a sphere or any such, but not to remesh.
he talks about the 'flipping' in the second micromesh video
http://pixologic.com/zclassroom/homeroom/lesson/military-assets-with-joseph-drust/
Spliting the mesh up could help (something I usually do with micromesh geometry out of habit, since it can skyrocket the vertex count faster than a subdivision). Also consider the portion of the model you're sculpting; sometimes noisemaker might produce a result that serves well enough, or an insert mesh.
And for the sake of interest, got it to this stage, but the rings are too small, a step down and they're too big, hah.