@pascal
thanks; really helpful!
My guess is how obvious the effect is for both a physical part and a rendering depends on the lighting details andsurface characteristics, Also for a rendering the details of how lighting is simulated along with how reflections, diffusion, etc are calculated. Real world with a highly reflective finish under a a couple of spotlights in a dark room would look a lot more like an equivalent render (good renderer with sufficiently fine mesh) than would it would look like a another example of a physical part with a diffuse finish with very even, flat (no shadow) lighting.
The effect Tone complained about is still present in Jeffās rendering from Indigo but not nearly as obvious as the rendering in the first post.
How close depends on how skill, time and money is available to make the physical part as well as the particulars of the geometry.
It gets much less straightforward when the joint is not planarā¦
-Pascal
are there any tutorials/books/theory on this topic? iāve watched some of the lynda tutorials but they donāt even talk about continuity. There is the NURBS book but while interesting I doubt the practical helpfulness of knowing the exact math terms that define a eg. a g4 surface. I figured most alias modelers know this stuff but tutorials on alias are hard to find (/and of limited use for rhino users 
) any suggestions?
well, itās a different effectā¦ the render i posted shows the flat surface as a consistent tone which then changes once the surface direction changesā¦ (a tangent blend will show this more readily as itās an abrupt change from one direction to another where as a smoother type of blend eases into the transition.)
but with toneās original picture, it looks like the flat surface is sitting lower than pipesā¦ or if you ran your finger over the surface, it would dip into a depression even though thatās not how itās actually modeledā¦
i donāt know, it looks like that to me at leastā¦ surely it looks like that to everyone else as well, right?
(original picture)
(an attempt at trying to describe in words what i see when looking at the picture )
point beingā¦ i agree with what tone is sayingā¦ i donāt agree with the rage thoughā¦ no way to have a proper discussion if coming in hot like that 
so to discard any doubts that this is a rendermesh or rhino issue there would be only one solution: render it without a rendermesh meaning rendering directly agains the nurbs surface, right? One of the programs that can do this is keyshot 5 pro so here you go: 
also the effect is partly a optical illusion; the center patch is of constant value
that actually shows the problemā¦ the center patch should be of constant value which your analysis is showing itās notā¦ itās flat and should have an even tone throughoutā¦ your picture shows that the tone isnāt remaining consistent at the edgesā¦ itās putting unrealistic highlights and shadows at the edges of the surfaceā¦ Vray, keyshot, most renderers are going to do thisā¦ Maxwell, indigo, and Thea at proper settings (unbiased) will probably do it correctly thoughā¦ those renderers donāt try to fake it for the sake of speedā¦ itās the whole āmaxwell takes longer but the quality is betterā argumentā¦ this is an example of what they mean when saying that.
(and lighting, materials, etc are going to have different effects as wellā¦ as already mentioned by others)
I agree with Jeff that the object in the original render appears to be depressed in the center flat area. Itās a good example of how the eye and brain can interpret or mis-interpret āedgesā in images.
1 Like
@jeff_hammond well, did you even look at the second rendering? the center patch has constant value and:(you have to click on the image for the animated gif)
yeah, thatās what i was referring to in my postā¦ and itās doing the same thing rhinoās rendering does. it makes it look like the flat surface is sitting lower than the pipesā¦ a real world object in real lighting isnāt going to have that phenomena(?) except under very exact/rare lighting&material circumstances.
i donāt consider myself a rendering software expert or anything like that and i hope iām not coming across as saying suchā¦ iām just trying to say that when i look at the picture in the original post or your keyshot render, the flat surface appears to be dipping in or lower than itās modeledā¦ does it look that way to you too?
yes thats the way it looks to me aswell; but: do you see the same way wich the greycolor around it? or in the animated gif? I dont. I assume its a optical illusion and when you check the colorvalue in photoshop the gradient is linear (better without bumps) and there is simply no dark edge ; please open the file and zoom to 3200% and you will see.
Hi, Jeff
I think youāve described the problem perfectly. The point I was trying to make earlier is that this is common lighting problem. Not exclusive to Rhino. It shows up with every Cad package Iāve tried. The built in screen lighting in all of them will display a dip. You can disguise it with a good quality rendering with shinny materials. Matte materials for some reason will still display a dip as seen in @eleven10 renderings . Maybe itās something to do with overhead lighting and surface edges. It doesnāt appear to be affected by the mesh settings.
ohā¦ didnāt realize itās an animated gifā¦ isnāt displaying as such for me automatically but if i expand it, i can see what you did there.
yeah, it could very well be an optical illusionā¦ afterall, weāre trying to look at 3D objects on 2D panel using who-knows-what amount of simulations/approximations/etcā¦ surely something will be lost in translation between that and real world objects.
just fwiw, i donāt have a problem with rhinoās rendered view and iāll often simply use screenshots of the viewport in presentationsā¦ itās way sweeter than what iām coming from in my previous app (sketchup)ā¦
still, itās an interesting enough topic to try to come to some sort of conclusion on.
Oh well, I was thinking about that matter, and after digging with Gimp, I came to the same conclusion: it is not Rhino which is going wrong but our perception. To add some more, not only the triangle is of consistent colour, but even what we see as a highlight near the borders of the edges, it is not actually there, e.g. colour changes in a straight forward way, no up an down, only up or only down. A funny point is also that I was going to post the picture with the chess table my self before seeing your post.
Regarding the continuity in y pipes, my believe so far is that it is not possible to have other than tangency - g2, because we have to deal with circular ones, as in the current example. If we would like to go in say g3, I suppose we would have to have a transition surface, like a collar on each pipe, which will be g3 capable but not circular. In that case, I believe that, from a creative perspective, it would be more sincere - and also more intereseting - to implement a completely seperaqte middle part
greets
yepā¦ iām able to get the look in Indigo if i use Sun for the lightingā¦ (i used a studio hdr in the earlier example)
(excuse the firefliesā¦ only 30seconds)
so yeah, this is probably a case of the limits of our current technology as a wholeā¦? (or jeez- maybe the thing actually would look like that if i 3d printed it an set it out in the sun?? and iām wrongly assuming it wouldnāt? anybody got a 3D printer on hand? or other means to make a test object? 
)
2 Likes
Hey, you are lucky because for some unknown reason even the first file was loading in 4 Ok , I am joking, but also I am not sure about your model and the subject of blending. Do you mean that you or anyone else would ever go on and fillet all the rough edges of a building? Oh no, I would not do that even to my worst enemy All right, I could only fillet some parts if I could have to show up close, but for general views there is no need to do so, I believe, and if I remember correctly there are some renderers (V ray maybe), which add some softness in rough edges at render time. But I can not tell you for sure
greets
ha. thatās crazy.
The dip wonāt be there with the manufactured part. However the sliver points of the planar surface will be very noticeable. Thatās why itās best to avoid this type of topology, I would recommend @pascal multiblend approach (easier and better with VSR) or @anon66739973 method.
I really hope weāve thoroughly exhausted this topic now.