One smooth surface made of stairs and outdour chair

hello everyone , im trying to create a smooth continous chair surface that has a staircase on one corner , i couldnt figure out how , i already built the constructive curves but whenever i seem to loft i face some errors , i want the surface to follow the 1st picture but i want it one smooth final surface (2nd picture which has some errors) please any have some insights? thanks!

3D MODEL_02.3dm (8.3 MB)


i think you need to more precisely define what you mean by “smooth”.
what should happen along those sections ?
kind regards -tom

it should be one signle continuous surface just like this , if you open the file i sent there are already the constructuve curves that show how i want the result to be! , the curves all already have the same amount of control points!

Rhino is a very precise tool, the less you care the less it will work out for you.

achieving smooth results when you have such input geometry might be challenging. better clean that up before you proceed.

start with levelling these curves


also your curves are not very continuos.

when you clean it a bit for a quick smooth surfaces you can try converting the Surfaces into a Mesh then use Quadremesh with option SubD. You can try using Quadremesh without meshing it first but i used your wonky curves as input and that might have made some hick ups so i had to make a tight mesh first and then use Quadremesh.

below is just a quick test, you might have to invest a bit more effort than that but it might work.

if you want to have more control you can try to fillet all the edges and stay within nurbs.

Agree with @encephalon quality curves > quality NURBS surfaces > fillet between the surfaces > done. It’s really that simple.

thanks but is this result a mesh or surface?can you also please send me the file? thanks!

its mesh, and its not very clean, you can see on the right side at the lowers stair that there is quite some distortion still, it would need quite some clean up and maybe a different mesh topology so you better start from scratch. it was just an example how it could look when you convert the Mesh or the Surface into SubD which inherently results in smooth edges depending on the amount of faces.

if that is a pool for production you might want to stay within Nurbs surfaces and create clean curves first then with the resulting surfaces start filleting the edges with either FilletSrf or BlendEdge etc. it might be a good practice for someone getting started. if you get stuck since it can be challenging merging several corners like here you could try Rhino 9 which has some refreshed kung fu in those matters.

Very true.

Rather, filleting between surface pairs ; ) which is an important differentiation, as @Jim pointed out countless times, where allegedly “impossible filleting” was well possible, if one forgets about edges and only thinks about the surfaces instead : )

you can get your current curves to work, Ideally I’d clean them up as others have said, but you can make gold from garbage if you have to- (sometimes you have to just make do with what you have)

loft your curves together in pairs to make “strips”

then use split by isocurve to split them back a bit top and bottom to create “gaps” between these strips that you can then fill with blendsrf set to tangent. You will need to add shapes and adjust the profiles of your shapes to get the blends you want.

join it all up and you have a result.

now, is this super refined and production worthy? God no…

does it smash something together for a concept, print, quick look model or costing scenario? yep.

Does it also very quickly give you the basis for something to refine and make “more perfecter?” at a later date? yep.

Ironically, this was very similar to one of the tutorials we used to teach at tsplines back in the day-

It is a perfect subd project, made in very much the same way strips, gaps, bridge, stitch, etc…

To make it more obvious, why booleans, joined surfaces (apart from final parts), and FilletEdge are bad ideas (as far as 3D NURBS surface modelling is concerned), because they lead the novice to “edge thinking”, is this example from Jim’s “filleting treasure trove”.

FilletEdge doesn’t “know” that surface C even exists.


yep, the strip and blend method tends to work well for these model where the fillet edge style model is a nightmare.

the key is to make sure the gaps between the strips are big enough to allow the blends to bend far enough to keep the shapes you want with the flat parts where you need them.

you can even blend to a blend once you get going.

There’s a good reason why clay and foam modellers learn where to put tape before introducing fillets/blends between the overbuilt main surfaces ; )

Once you show industrial and transportation design students how to manually make a G1 or G2 fillet on a piece of timber, foam, or clay, and then let them do it in Rhino (or “other software”), it clicks, and “edge thinking” won’t infect them : )