Good critic. You are right. Thanks for the input. 
Always wonder where did you get that nice reflection materials. Some of them look like Alias onesā¦
They donāt look like Auxpeckerāsā¦
One last change I would make is to merge the two side surfaces with smooth off before trimming. That way you remove the pesky sliver point which can cause problems.
I donāt use auxpecker. I donāt like any of the display options. I create my own with lighting studios and abstract images while playing around with these settings. 
1 Like
Hello Tone,
May I ask what is the purpose of the model? I mean, is it going for physical production or for visualisation purposes only? I am saying so because you are absolutely correct about the title of your topic - Render Rage. Rhino is failing hard on showing this type of surface continuity (tangent), so I would suggest to skip all of the answers here if your model is going for production. Or skip Rhino all together if you are going for visualisation purposes - unless of course there is a renderer for Rhino able to sensibly show off the surfaces.
And now on to my solution. I corrected the model in some ways, as while the intentions of the design were clear, there were some minor bugs; e.g. some of your quarter circles were of degree 3, while 2 is sufficient. Also, the surfaces were not really tangent, indeed, all thought as I pointed out, the intentions are correct. So I ended up re implementing the surfaces on a new layer, on top of your previous model. I guess you can see the logic behind the modeling I did; I highlighted the arc centers, used for the blend surfaces, made up of revolving the arcs around those points.
y pipe custom.3dm (86.7 KB)
P.S. Please, if you wish, could you mind save your work on Rhino 4 compatible format; in this case, I suppose more people could be able to help you - for example, I am on Rhino 4 and I am running out of saves in 5 demo. Also, āSave smallā will not include the mesh in your file, so from 5 MB my file went down to 86 KB
Edit:
I took the āfreedomā to shave off one of your cylinders, in order to have a circular arc blend, so please compare the ācustomā layer with your original file if you would like to see the differences. I am not sure how things go on with other than circular arcs so what I presented here may not be very well suited for your needs
Greets
Hello Demosthenes_Stogios
I have never been able to blend these kinds of shapes successfully. When I arrived at a solution I was satisfied with only to discover how badly Rhino renders tangent flat to curve surfaces, I had a rant. I donāt use Rhino in a professional capacity (nor would I ever wish to) but I do like to make things in it from time to time. Usually I do architectural stuff but lately I have been dabbling with old machinery and am coming across these blending issues. I attach a poly model of Paladioās Rotunda to show the lack of blending I usually get up to.ROTUNDA to vue.3DM (3.2 MB)
Since you @Tone @Demosthenes_Stogios are convinced this is a failing on Rhinoās part. Please provide screen caps done with other Cad modelers of this part that are significantly better.
Thanks
3 Likes
Just to add my 2 piasters worth for the sake of variety - hereās yet another way to go about this- not exactly automatic, but reflections fall cleanly in the transition area, which is nice - no flat spot and no dips.
y pipe_PG.3dm (192.3 KB)
1 Like
@pascal great solution!
would you mind sharing a bid of your process?
esp. how did you construct the middle point? just the intersection point for all middle isocurves?
how would position it if they would not meet (like in a asymmetrical 4 sided patch)
also in which order and how do you construct and match the tertiary srf.?
Hi Konrad -
in this case it is easy to locate that point since the joint is planar - finding something close to it in more complex situations might be a matter of some guesswork and luck I would say.
That point sort of drives the rest - I split the existing edges somewhat arbitrarily into thirds - Iām not sure that is ideal but it seems at least OK.
- Blend curves between the lower edges of the half-pipes
- I put an arc (copy of the 1/3 arc from one of the existing pipe edges normal to the blend curve at the middle point location.
Well, hereās a little clip, maybe that is easier - Y-Joint- round tubes.
The rest is careful MatchSrf, mostly for tangency and it gets pretty close to curvature.
-Pascal
5 Likes
nice!
you should talk in thoseā¦ just be you. theyād be great.
3 Likes
I believe I have been clear that I was commenting on the rendering of a flat surface tangent to a curved one, not the modelling of it. I couldnāt care less how many modelling packages out there show the same thing. If you believe you can model such a surface combination in Rhino and render it in Rhino and not have it looking like it was stamped into the surface, show me.
My point is it will look the same no matter the Cad program. Itās in the nature of attaching degree 2 surfaces to planar ones. You can only achieve your desired result with degree 3 surfaces with a G2 match. Hereās Demosthenes model exported as a stp file into solidworks. I maxed out the display settings, solidworks aliasing sucks. 
Hello Stratosfear
I am not that technically minded when it comes to surfacing issues, as you have probably gathered but may be if I come at it this way: We never see the nurbs surfaces only a polygonal representation of them (the render-mesh). I believe the problem lies with the render-mesh; I donāt believe polygons care about G0, G1, G2 etc (though I may be wrong). Sorry if my previous reply was a bit snippy.
No, the problem is the surfaces are tangent and not curvature continuous. Several examples have been provided that are G2, curvature continuous, that do not show the tangent to flat āproblemā.
-Pascal
The rendering is calculated from the geometry of the rendering mesh, not directly from the NURBS surface. But the geometry of the rendering mesh depends directly on the NURBS surface. The orientation of the polygons in the rendering mesh will change between adjoining polygons more smoothly along the edge between surfaces if the surfaces are G2/curvature continuous than if the surfaces are G1/tangent continuous. So the polygons do care about the continuity between surfaces. The result of the less smooth change in polygon orientation across the edge results in the visual effect you see. If you made a physical part with a smooth surface with the exact shape of the NURBS surface you would see the same effect.
while the same effect would occur, I donāt think it would be the exact same. itād be much less pronounced. (talking about flat ā tangent transition)
Why would it be much less pronounced, assuming a fine render mesh was used?
because the lighting would be realisticā¦ real light on the object wonāt look like toneās original post.
a quick example:
using this mesh from rhino: (toneās original .3dm upload)
ā¦ in an unbiased renderer (indigo)ā¦
ā¦thatās a lot closer to how the part would appear under real world lighting conditionsā¦
that said, i donāt think weāre there yet as far as what type of lighting can be shown real timeā¦ the render took 4 minutes (and would take longer with more complex materials/light)ā¦ i think rhino uses openGL stuff in order to display realtime in the viewport rendered modeā¦ itās not perfect but itās a good compromise between speed & realism
Also, Iād guess that on a part made of real material - bent metal or what have you, there is no such thing as strict tangency as there is with digits.
-Pascal