Pro v4.2.2 - I find that I need to wall some vectors and used it successfully on some path nodes, and had it fail on others. Instead of the wall using the specified axis, it skews the faces out sideways.
Undaunted, I thought I could use Loft instead, but no - loft fails and the staus log reports I cannot loft with just 2 points.
OK, so I add a vertex to each edge, and loft works.
Hm.. So I try walling the vectors that failed originally (but with the inserted vertices) and hey! That works!
So my REF is to ask if this limitation can be removed and allow us to loft or wall 2 point faces (given there is an axis specified I don’t see why it should be a problem)? Thanks!
In the meanwhile I will be adding vertices to my 2-point vectors before wall-ing.Hm - it seems the inserted vert doesn’t fix the wall function as the new faces are still skewing sideways instead of “up” along the specified axis.
Pro v4.2.2 - I find that I need to wall some vectors and used it successfully on some path nodes, and had it fail on others. Instead of the wall using the specified axis, it skews the faces out sideways.
First (sorry for the nitpicky terminology correction, but…), you’re not really walling “vectors”, you are walling polygons.
Next, wall requires polygons that have defined “normals”. Polygons with just 1or 2 vertices never have defined normals. Polygons with 3 or more *usually* have defined normals *as long as the polygon is planar and has “some” area*.
Finally, these seems like reasonable requests (for both Loft and Wall). Note, however, that if Creator allowed you wall 2-point polygons you would not get to choose Direction:Face Normal (because the 2 point polygon does not have a normal). It would still allow you to choose other Direction types: X, Y, Z axis, etc.
Fair enough Steve, that should have read “path nodes” which are in reality wireframe polygons. The limitations you noted are acceptable for me since the global axes are available to be specified.
I presume that any local coordinate system in use would also be supported..?!
FYI it looks like there is a bug in the wall tool. I was thinking all I had to do was insert vertices into a 2 point poly et voila, wall would work, but it turns out if the wireframe has vertex normals and they are not 0,0,1 then the walled faces skew sideways (I know it worked at least once but am baffled why it comes and goes - I did modify my edit settings in between, maybe the answer lies in there).
However, I can set the wireframe vert normals to 0,0,1 and they wall “up” fine. I need to wall along the Z axis so this works; but it’s interesting that even though I set Z Axis in the wall dialog, the unmodified faces wall sideways.
The Wall tool will only work—reliably—on a polygon if the normal for the polygon is defined (i.e., can be calculated). As I mentioned above the normal can only be calculated for a polygon that contains 3 (or more) non-colinear, coplanar points. The normal calculated for such a polygon is the vector normal to the plane in which all the verts of the polygon lie.
Vertex normals have nothing to do with polygon normals - these are completely separate. So in your case, Shawn, if you simply add a vertex between the two existing points of a 2-vertex polygon, you still have a polygon composed of 3 colinear points and that polygon won’t have a “defined” (calculable) normal and you will get undefined behavior in the Wall tool. I think that is what you are seeing. It’s not that the Wall tool has a bug but you are feeding the wall tool a polygon for which the normal is ambiguous. The problem with such a polygon is that there are an infinite number of planes containing the polygon so the normal vector cannot be reliably calculated.
And if you wonder why vertex normals don’t matter to the wall tool, imagine a polygon that has vertices whose vertex normals are all different. What normal should the tool use?
OK, I can’t let it go like that - see the images for how the tool fails for a 2 point face, but works OK for one that has a vertex inserted. I don’t know/care how that bolsters the argument that the tool does/doesn’t work for colinear faces - to be honest, I just want what I want
It’s confusing to me why it fails sometimes - as you can see by the undoctored images, it does work OK with some colinear constructions.
I’ll work around the failures as they happen. NOW I’m moving on
Sorry if I misunderstood your post in which you claimed the tool behaved differently based on the values you set on the vertex normals. My point is that the Wall tool ignores vertex normals and only considers polygon normals - so I suspect your observation that the vertex normals affects the behavior is a red herring.
I’ll let you get back to work but if you want to follow up, please clarify the problem so we understand each other.
Well - Maybe the conversation got a bit derailed over vector vs face?
I thought I was clearly saying my observation of the tool failing to work was based on vertex normals - apologies if that was not clear. I recognize sometimes I look for ways to use the tools that they may not be intended for - I just need to get things done! I am using wall on a wireframe poly to create a 3D shape in order to boolean against another 3D shape to slice it, and the wall function (loft where wall fails completely) is a required step. I care about the curved portion of the walled faces and where they slice the 2nd 3D object.
The last example documented above is missing one step (attachments are limited to 3 total). I in fact inserted one vertex along a 2-point edge and the wall function failed (even though this usually works). When I inserted 2 more into these new edges (2 just to be on the safe side), then the wall function worked, and I used that snapshot for #3. These points are ALL colinear, since the starting point is a 2 point face (wireframe) every time I have to insert verts.
I think this is what happened before when I said it didn’t work - and since the additional verts are all colinear, it makes no sense to me why there would be any difference between adding 1 point vs adding 3+ (:shrug:)
The last example documented above is missing one step (attachments are limited to 3 total). I in fact inserted one vertex along a 2-point edge and the wall function failed (even though this usually works). When I inserted 2 more into these new edges (2 just to be on the safe side), then the wall function worked, and I used that snapshot for #3. These points are ALL colinear, since the starting point is a 2 point face (wireframe) every time I have to insert verts.(:shrug:)
After doing some experiments, I noticed that Add Vertex—sometimes—does not work properly on a 2 point unclosed wireframe polygon. I’m still trying to investigate why this is happening but in—some—cases I noticed that after I Add Vertex to a 2-point unclosed wireframe polygon, the resulting polygon had 4 vertices (2 new vertices). One new vertex is put where you would expect it (between the 2 vertices) and another vertex is put after the the last of the original 2 vertices. The resulting polygon has 4 vertices and “doubles back” on itself. Even though the 4 vertices are collinear, the “double back” vertex tricks the Wall tool into thinking it can calculate a normal vector.
I think that is what you have been seeing. You can verify this by:
The next time you see Wall work on an unclosed wireframe poly that appears to have 3 collinear vertices, check for the 4th double back vertex. I beileve this will vindicate the Wall tool and shift focus to potential problems in Add Vertex (on unclosed wire faces)
Interesting - I never thought add vertex would be the culprit!
Shawn, did you actually verify this? Just curious because the problem seems a bit random on my side. If you see the same problem, I’d be more confident that this IS the problem.
Thanks in advance.
I didn’t verify after your post, but that’s consistant with some of what I was seeing prior to - why sometimes inserting vertex had no effect on the wall function. I will be doing more of this type of operation later today and I’ll report back.
That doesn’t matter eh?
