Additions to Fillet node
I am working on a personal project where I am generating a racing track from a simple vector-handled bezier curve. And as You guessed, that might require using a fillet node. But in my case, it has limitations, shortages, and lacks effective use. So my suggestions are related to them:
-
Selection of control points for filleting. Of course, I can give 0 radii for the point, but a simple selection looks like a typical way to go for such kinds of functions/nodes.
-
"Limit radius" option is not working properly as it should, because while limiting radiuses, which are becoming too large, it also involves decreasing smaller radiuses at neighbor points. That takes away control, logic, and precision of the outcome.
My guess is that radiuses should be limited by prioritizing the smallest defined not to be touched. I hope it is possible to make it at least an additional option or with the drop-down choice of algorithms for the radius limitation. Besides, "limit radius" boolean input does not allow to use for each point, but on the whole curve.
-
In my case, if I use a fillet node for the procedural modeling logic, I am really lacking the isolation of actual fillets. Same as in extrude node, the user would like to have an option to work with fillets or undeformed lines. If the selection of curve segments is not possible, probably output would include actual fillet and undeformed splines separately.
-
If the "adaptive" radius option is activated, it would be also useful to find out the resulting radiuses and their centers.
Use cases? - Fillets might have a separate tree/group with a special approach. Detection of resulting radiuses and centers can help in adapting bank angles/tilt, and profiles. In my project, I am then separating sections of the curve for sharp turns, less sharp turns, straights, and even trimming sections of the straights near to turns. At the moment, those lacks decrease the precision of the model and require huge tricky trees for detecting necessary data/features, especially involving braking curves down to meshes and then back again. At the moment, fillets are isolated by detecting the angle between edges after they are converted into a mesh. I think You agree that this is not how it should be.
Thanks to all the developers for making the blender and the geometry nodes possible. I have waited for such an option for a long time. Thank You!
1 Comment
Since they must already be calculated, a further output of 'Fillet Centers' would also be without significant cost, and very useful? (Ahh, OK, I see you've mentioned , but not illustrated that.