Thanks for understanding, since we can add a vertice to an edge without splitting up the adjacent faces, then I think we should have an option for merging faces with out loosing vertices too… But I think this should be mergefaces job, and then delete can do what delete does now.
delete would be faster, but not sure of the underlying code requirements there.
Here is more Just make two subD faces that share a single point.
1- Try lofting two edges that meet in one point (expect a tri here, but Rhino fails.)
2- Try bridging, this fails too.
So here we need to join them first and then append a new face.
This is a tedious trail and error workflow that takes focus away from simple tasks, so please see if 1 and 2 can be fixed, or at least give a nice feedback on why they failed.
I had ignored the sweep1 command as I thought I could use the ordinary one, so thanks for pointing out that it is a separate sweep1 SubD tool.
3 I often add single faces to a model, many after each other, so I also ‘need’ an AppendMultiple option where I can continue to append faces to the same mesh with out having to select the mesh again each time. (start command, draw face, hit enter, draw a new, hit enter etc and one more enter to end)
4 - Offset to work on subselections… or is there another tool to replace faces with their ofsets that builds the edge automatically?
Extrude could also do this, but neither WCS nor UVN gives this result. Or have I missed something?
Hi Holo, I also don’t like that the SubD bar hides the main tool bar, where I have hide, ect. I am going to see, if I can copy my whole left toolbar, where I keep the goodies.
[OT, but Nice truck! Ironically, my friend was doing a body-off-frame restoration on a truck pretty much like that.]
Edit: Good thread. Yes, make Rhino’s SubD powerful.
I thought of a rather strange feature that solves challenge with subd.
I want to be able to mark SubD polygons as either active/solid or wireframe/structural reference. In other words, to make edges of SuD assemblies that fit, you need overlapping/redundant edge areas, which we don’t want for rendering or other things, yet are required to keep the shape to fit other parts.
This would be the good first step toward trimmed and intersected SubD geometry, as one SubD could split/bisect another and mark overlapping/intersections, as non-solid, reference–just the same as it does for trimmed NURBS!
In the screenshot below, the separate SubD’s in the window opening fit well–only because they have overlapping/coplanar edges.
Nice Brenda, thanks for chiming in!
I didn’t fully understand the overlapping, non-solid, reference thing… do you mean trimmed-subD? If so that would be cool… but maybe difficult to maintain though.
I too model parts that shall fit with the same number of vertices in the same spots.
Thanks Kyle, my bad on the feedback, and thanks for the gumball tip, I’ll check it out. (Sweep1 with chained rail would be great too, at least I have not been able to use that)
And inset is great, but not if I want an offset, then I have to offset and delete the parts I don’t want and bridge.
Yes, Holo, the effect would be fairly close to trimmed SubD’s, the ability to mark SubD polygons to hold the shape only. (Thanks for having me on your thread.)
In other words, with traditional SubD, you just can’t split something into two–without changing its shape. I say, keep polygons that are needed to keep the shape, and mark not to be filled for rendering or any other purpose. The effect would be a little like trimmed polysurfaces, but at the SubD polygon level. Each whole polygon in the SubD, would be marked as a surface–or a shape-holder.
I feel that would help on the doorpanels and the headliner on your truck : )
Currently, I copy SubD’s, extract edge perimeter polygons (rings), and build the second structure from that. The result, unfortunately, is coplanar surfaces.
[Going much, much further, perhaps two SubD meshes could be intersected and not trimmed, but the subtract-e could be besected along the intersecting line, adding polygons to approximate a split.]
I have also wanted a no-draw material to speed real-time previews, (on my aging, yet mean/average GTX 1080) though, there may be advantages also, to being able to mark certain polygons of a SubD mesh as for-shape-only. Giving my druthers, of course I would want both features. : )
[I will point out, that a friend had said that Rhino does for some reason, already hold its edges in a more beneficial way than Modo’s, this would be another extension of that. This as related from someone who used Modo for one of the largest advertising accounts, fullstop.]
I see what you did there, nice sell in!
Regarding speed-up. Maybe we could have a per-object subdivision override.
Now it uses a 8x8 (3 times subdivide) for each face, no matter the complexity or size of the object (as far as I know Rhino doesn’t support LOD (level of detail based on distance from camera or size related to pixels the object occupies))
If I extract the rendermesh from half the car I get 900.000 polygons… that’s almost 2M for the entire car at this detail level. So it requires quite a lot of computing power for the GPU (even though it’s silky smooth on the 2060 I have. I get 220 FPS at rendered mode at quarter screen size (Ultra wide monitor, so the viewports are 1470x770 each) (Test with -TestMaxSpeed and set VerticalSync=Off, then run twice)
subd loft tune up bug track here-
https://mcneel.myjetbrains.com/youtrack/issue/RH-68062
thanks all for this stuff… it’s helps a lot to have such great feedback.
Right back at you Kyle! Getting response, hints, tricks and exchanging ideas is rewarding and makes it all worth while. And having a hobby project where time is no limit makes it possible to report stuff too. Not too many customers are willing to pay for that during projects… come to think of it… none are… but some are none the less, unwillingly
Did some more tweaks an added a few glossy details. All subd of course as these builds are purist stuff for the fun of it. Oh, and threw a PBR material on it too to get the car paint shine, hope you like the color, deep red is always great together with, crome, glass and it’s shadowy areas IMO. All lit with a single HDRI as it’s a much better modelling light than panels, so I think I’ll use that for the final renders too, if I ever get to that stage.
good lord…this is looking amazing!!!
one nit pick with most subd cars… the panel gaps are always a little sloppy… (you have done way better than most)
You may consider merging all the panels (hood, doors, etc) into continuous surfaces and then converting to nurbs at the end, and then add in your panel gaps with a nurbs trim.
or try creasing the edges, then convert to nurbs and fillet.
or crease the edges, add tiny bevels and then delete the creases to keep it all subd .
Or ignore me completely and carry on with this awesome model.
Thanks!
Yeah, I agree and I keep them single faceted while I get the shape right and then I add an extra face to the edge when done. Too high complexity while tuning is just too much work.
Does this look better to you? Hood transaction only. (…still lot of fine tuning to do )
chef’s kiss<
Is there a way to remove an edge bevel where the adjacent faces keeps their planes? Kind of like fillet curve with 0 as value does? So we can add an edge ‘fillet’ and then remove it again?
at the moment no-
you have to do that manually