LDraw.org Discussion Forums

(2019-06-06, 9:53)Philippe Hurbain Wrote: [ -> ]Yuck... Can you share the files?

Only late evening, I started a render on my laptop which turned out to use more energy than anticipated... 

I will also try with other transparent parts that consist of sub parts.

Here are the files, I checked if I may have set the wrong part-type, but all are sub-parts...

I just tried to render 10312, the big bubble canopy which I did as well and I split also the top surface into four. I get the same seams...


[attachment=3731]

I tried to inline the part and the result got worse:
[attachment=3732]

A last try was to have the original generated from the CSG tool and take it "as is" (plus cond-lines), that looked quite perfect:
[attachment=3733]


Dou you think it has to do with the quad winding when the sub-part is inserted and mirrored? That is the only thing I can imagine

(2019-06-06, 20:58)Gerald Lasser Wrote: [ -> ]Dou you think it has to do with the quad winding when the sub-part is inserted and mirrored? That is the only thing I can imagine

I think it might has something to do with Z-Sorting as the seems are visible in LDCad too, caused by 'double' alpha blending of certain polygons.

LDCad does part level z-sorting, while LDView does polygon level sorting hence no seam.

Don't know how that leaks into ray tracing though, as I thought Z-Sorting was a rasterizer specific thing.

(2019-06-06, 22:08)Roland Melkert Wrote: [ -> ]I think it might has something to do with Z-Sorting as the seems are visible in LDCad too, caused by 'double' alpha blending of certain polygons.

LDCad does part level z-sorting, while LDView does polygon level sorting hence no seam.

Don't know how that leaks into ray tracing though, as I thought Z-Sorting was a rasterizer specific thing.

Studio does rendering with the GPU, not ray tracing.

Weird problem. Looks as if subparts (or primitives!) are somehow rendered separately.
I think you made a wrong manipulation for inlining, here it works perfectly.
In the following image, left is a triangle soup (deep inlining), middle uses 1-8sphe primitives, right uses 4 subparts with inner/outer skin.
[attachment=3734]
Studio POV render looks good:
[attachment=3735]

Thansk Philo for your experiments!

It is a pity that eyesight is doing it like this. Can we ask BL/Stud.io team to look into this or talk to Eyesight? Shall we just open a thread in the forum over there?

A lot of the bubble canopies are using the approach of having the surface cut.

(2019-06-07, 12:10)Gerald Lasser Wrote: [ -> ]It is a pity that eyesight is doing it like this. Can we ask BL/Stud.io team to look into this or talk to Eyesight? Shall we just open a thread in the forum over there?
A lot of the bubble canopies are using the approach of having the surface cut.

Yes I think it could be a good idea to tell about the problem on Studio forum. But it's kind of hard to get attention there: eg. the shortcut rendering problem reported by several people (including myself!) didn't get any feedback (maybe I missed an answer, anyway the issue is still there). This problem is imho even more annoying, see https://forums.ldraw.org/thread-22940-po...l#pid29473. Corrected image by inlining here: https://forums.ldraw.org/thread-22940-po...l#pid29486

I have completed two more crazy prints...

Cleaning in progress, Cockpit from Ant-Man and Lego Movie 2 shall be up soon!

[attachment=3799]

I noticed a nice part recently, the ornamented fence got a quarter circle sibling!

... and I realized only when posting this, that it is 3 x 3 only and not 4 x 4 like the other quarter circle fences.

So I played around with adding cuts into the straight piece and rotating it on the 4x4 base, which turned out pretty ok...until I, well , realized it is onyl 3x3
[attachment=3847]

So anybody has got this part already or has got pictures from below?

May be this is a sign of a 3x3 macaroni brick coming?