if i place one surface (of a group/dynamic component) on another surface, so that they are level, they are visible half-half, someway randomly.
What is the rule here, that looks random to me? (How) can i get around it?
More specifically, I model a brick wall, and for low detail plans, i just cut the clear dimension. Now when i add more detail, aka the wooden window frame, i want that frame precede the wall in visibility:
I want to keep the wall as-is, so i can still get low detail plans from it, just by disabling the layer with the window frames. So Intersecting the frame with the model and deleting the affected parts of the wall is not really an option.
A painful way would be to intersect, hide the surfaces when showing the frame, and hide the edges when showing the wall only… there must be a better way, right?
This is called z-fighting. There have been many discussions about z-fighting and ways around it here on the forum, some quite recent. I suggest you use the Search feature of the forum.
The short answer, though, to the Rule of Precedence question is there is no rule of precedence other than the same rules you’re already familiar with from Earth, namely, if Face A is between Face B and the camera (your eye), Face A is in front. If Face B is between the camera and Face A, Face B is in front. When two faces are congruent, neither face is in front, and z-fighting occurs.
The practical solution in this case would be to treat the window frame as it would exist in real world terms. The frame would not occur on a plane congruent with the adjacent wall, but rather, would be slightly removed from the wall surface. Modeling the frame as a separate entity attached to the wall eliminates the z-fighting issue.
Thanks, i read some on z-fighting now. The problem in principle was clear to me, but the term was new to me.
I hoped for a property/setting for components that would allow to “overrule existing surfaces” or so…
As for the practical solution in my case, the wooden frame IS on one plane with the rock wall. But i do not want to cut the frame’s space out of the wall, so i can still create low detail plans from the model, by hiding the frames.
I described one complicated solution above, but now i have a 2nd idea, maybe easyer to do:
Make the frame say… 0.5mm stick out of the wall, small enough not to be visible in Scenes/LayOut docs (the rest of the model is done with 1cm resolution). Will i need to set a higher resolution temporarly, or do i need some other trick?
Or does anyone have other ideas on how to do lower and higher detail scenes (and from them LayOut docs) in one model?
If you model objects as they exist in the real world–where materials have thickness and two objects can’t occupy the same space simultaneously–you should rarely run into problems with z-fighting.
As for the various workarounds–self-cutting openings, scenes using layer visibility, hiding certain strategic edges-- these are discussed rather at length on this forum. I get the feeling you haven’t really read through that stuff, which, I suppose is no skin off my nose.
I have no idea what “1 cm resolution” means. I have no idea what “overrule existing surfaces” means.
You would be wise to get your arms around some of the solutions to z-fighting experienced users have proposed.
Thanks gully.
I thought i explained well why i think i can/should not model after the real-world, and i asked for other ideas on how to approach that.
With 1cm resolution i meant… what i think is called units>resolution in english (using german). I am not sure if there will be problems later doing what i supposed. One problems could be measurements in LayOut later, when i use slightly different dimensions just to overcome a presentation problem. Sounds like a bad hack in a way, but maybe it could work for my case…
With “overrule existing surfaces” i was trying to name a hypothetical feature, that would resemble something like layers for visibility, which SketchUp does not have. Something to tell sketchup what to do, if there are two surfaces in the same place…
Like a checkbox on a component’s properties that says “show component’s surfaces in front of ones from the surrending model”
or bluntly “let component’s surfaces win in z-fighting situations”
I DID read quite a bit, but all cases were more like “why am i seeing this flickering” and nothing that i saw a way to adopt to my multiple detail-level problem…
I guess if there is no good workaround, i will ditch the option to output low-detail plans and model after real-world.
If you don’t want to do as gully suggested, divide the face so you can paint the area of the window frame separately from the rest of the wall. Then you won’t have the z-fighting issue.
Invest some time in how “Glue to” plus “Cut opening” components work.
If you include a close loop of coplanar edges without a face in the detailed window component A and do the same with the simplistic window B and assign different layers to A and B you can switch between both versions without Z-fighting.
