Working with Rhino 6’s new rendered display is a dream come true!
But the clippingplane has an odd vertical offset that is too hight when working in Meters as unit:
2 Likes
We are aware of this and it’s being worked on (https://mcneel.myjetbrains.com/youtrack/issue/RH-42679).
1 Like
Can you do my a favor… Run “TestZBiasFactor” and click the “Less” option until the zbias is 0.0625…
That should clear it up…However, what I need to know is if doing that messes up anything else… So the favor is that you continue to run and work with the zbias factor set to 0.0625 (perhaps place it in your startup otptions) and see if you get any situations where wires are no longer showing through the render meshes…or if you get any kind of dotted pattern of wires on render meshes (wires dipping in and out of the render mesh).
I’m hoping that with V6’s new display engine(s), that the zbiasing isn’t as big of a problem as it has been in the past, and that I can just reduce the overall internal value we’re currently using… But I don’t want to do that if it’s going to break other situations and models/scenes.
Let me know,
Thanks,
-Jeff
1 Like
Hi Jeff, on a quick test here
or 0.03125 seems to be good.
I guess working on the terrain around a building in meters instead of millimeters are pushing your default settings 
Could the zBias settings be controlled by the files absoulte tolerance instead of it’s own settings? Or maybe even better: be controlled by the cameras distance to the focuspoint?
Well, it’s not just a number…it’s actually a bit complicated… we do take camera distances into account, as well as near and far clipping planes, and the separation between them, and a few other items in our algorithm in order to determine how much to nudge curves towards the camera… The problem is that it’s not perfect, and it doesn’t catch/work on all cases… Which is why TestZBiasFactor exists… The “zbias factor” is simply a “fudge amount” that gets applied to the result used in the cases that don’t work…it basically lets you tune the algorithm. The smaller the number you have to use means the better and more accurate the algorithm is. What is happening in V6 is that we’re getting better and more accurate results with the new engine, and the current fudge factor is now too much…there’s also a bug with clipping planes where the meshes are actually getting biased as well (I will fix that)…but that problem revealed a curve biasing issue, which is what I’m trying to address here.
What I’m basically after at this point are scenes that will not work using the small zbias factor (0.0625)… What you’re looking for are situations where you should see wires but don’t. You should not be seeing the opposite of that (seeing wires when you shouldn’t)…if you are seeing that situation, then it means the zbias factor needs to be even smaller. Did you try 0.03125 because you were seeing wires in cases where you shouldn’t? If so, can I get my hands on that file? I guess I’d like any file you have that doesn’t work (in either case).
Thanks,
-Jeff
Hi @Jeff,
I am uploading a file to your attention where the “default” doesn’t work (edges showing through).
By the way - I think the custom setting for TestZbiasFactor if set, should be somehow remembered per-file.
–jarek
1 Like
@Jarek Thanks, I got it.
However, it looks like a ZBias Factor of 0.0625 fixes things in this model. Is that not the case for you? I don’t think I was clear in my previous post… I want to see models that fail even after setting the TestZBiasFactor to 0.0625. I’m already aware that V6 is currently using too high of ZBias value as its default…and is what I’m working on fixing. When I load your model and zoom in to the balcony areas, I do see biasing errors…but when I set TestZBiasFactor=0.0625, everything looks fine here.
Is that not the case for you? If not, is there a ZBias value that does work?
Thanks,
-Jeff
Hi Jeff,
Yes, it works with the lower value well.
Sorry, I thought you needed any model that the default doesn’t work with!
–jarek
@Jarek No worries… Thanks for the model…it’s now part of the ZBias regression tests 
@Jarek, Sorry… I meant to say your model is “NOW part of the ZBias regressions”… talk about a single typo that changes the meaning completely…
-J
1 Like