Bevel by Offset produces inconsistent results on identical edges. #53459

New Issue

Matt Matteri · 2017-12-03T15:22:26+01:00

Matt Matteri commented

2017-12-03 15:22:26 +01:00

System Information
Windows 10 Pro
GTX 1080 Ti

Blender Version
Broken: 2.79 5bd8ac9

Short description of error
When beveling with the 'Offset' Amount Type:

Edges are both not beveled by the given offset but by some other amount.
Edges (even identical edges) are not beveled by the same offset.

Exact steps for others to reproduce the error
Create a symmetrical object
Select symmetrically identical edges and bevel them using the offset amount type
Observe result

Note: Even this is not consistent. I have created the exact same mesh, with vertices in the exact same positions, but different vertex indeces, and the edges bevel differently. beveldemo.blend

**System Information** Windows 10 Pro GTX 1080 Ti **Blender Version** Broken: 2.79 5bd8ac9 **Short description of error** When beveling with the 'Offset' Amount Type: 1. Edges are both not beveled by the given offset but by some other amount. 2. Edges (even identical edges) are not beveled by the same offset. **Exact steps for others to reproduce the error** Create a symmetrical object Select symmetrically identical edges and bevel them using the offset amount type Observe result Note: Even this is not consistent. I have created the exact same mesh, with vertices in the exact same positions, but different vertex indeces, and the edges bevel differently. [beveldemo.blend](https://archive.blender.org/developer/F1278418/beveldemo.blend)

Matt Matteri commented

2017-12-03 15:22:26 +01:00

Changed status to: 'Open'

Matt Matteri commented

2017-12-03 15:22:26 +01:00

Added subscriber: @Rekov

Sergey Sharybin commented

2017-12-04 10:19:14 +01:00

Added subscribers: @howardt, @Sergey

Howard Trickey was assigned by Sergey Sharybin

2017-12-04 10:19:14 +01:00

Sergey Sharybin commented

2017-12-04 10:19:15 +01:00

@howardt, mind having a look here? Thanks! :)

Howard Trickey commented

2017-12-04 13:38:49 +01:00

Yes, I see what you mean. I will look into it.

Matt Matteri commented

2017-12-04 18:15:49 +01:00

This bug might not be related to bevel specifically. Here is the discussion thread that led to this bug report, and it looks like the mesh itself might somehow have become corrupted?

There's a good chance something wonky is going on, though it might not be the bevel tool specifically.

This bug might not be related to bevel specifically. Here is [the discussion thread ](https://blenderartists.org/forum/showthread.php?441822-Why-does-bevelling-by-offset-produce-uneven-results-on-identical-edges) that led to this bug report, and it looks like the mesh itself might somehow have become corrupted? There's a good chance something wonky is going on, though it might not be the bevel tool specifically.

Howard Trickey commented

2017-12-05 16:45:10 +01:00

The mesh isn't corrupted. But it does hit a non-determinism in the current algorithm. Here is what I wrote on that thread, which shows my current understanding of this bug:

The problem here is that there are too many constraints that bevel has to satisfy simultaneously in this situation, and it is impossible to satisfy them all, so some compromises are made. Among the constraints are:

we would like the newly created vertices to slide along the existing non-beveled edges (users kind of expect this)
the slide distance should depends on the (sine of the) angle between the beveled edge and the slide-along edge -- because we want the perpendicular distance to the offset edge to satisfy the offset=amount spec of the user
there are two angles involved, and their sines may not be equal
all of these constraints hold at the other end of a beveled edge too, and again, the sines of the angles may be different; yet we would like the width of the offset strips to stay constant along their lengths to the extent possible

Because it is impossible to satisfy all these constraints simultaneously, there are compromise values chosen for how far to slide each vertex along their edges. Then there is an "adjustment" pass that tries to propagate slide the compromises to adjacent edges. That process unfortunately is non-deterministic - it depends on the original ordering of vertices to some extent, though there is an effort to try to chain the propagation together after initial choices are made among connected components of edges. Unfortunately in the model here, different choices of propagation order are chosen for two of the edge chains vs the other two, resulting in different end results.

The mesh isn't corrupted. But it does hit a non-determinism in the current algorithm. Here is what I wrote on that thread, which shows my current understanding of this bug: The problem here is that there are too many constraints that bevel has to satisfy simultaneously in this situation, and it is impossible to satisfy them all, so some compromises are made. Among the constraints are: - we would like the newly created vertices to slide along the existing non-beveled edges (users kind of expect this) - the slide distance should depends on the (sine of the) angle between the beveled edge and the slide-along edge -- because we want the perpendicular distance to the offset edge to satisfy the offset=amount spec of the user - there are two angles involved, and their sines may not be equal - all of these constraints hold at the other end of a beveled edge too, and again, the sines of the angles may be different; yet we would like the width of the offset strips to stay constant along their lengths to the extent possible Because it is impossible to satisfy all these constraints simultaneously, there are compromise values chosen for how far to slide each vertex along their edges. Then there is an "adjustment" pass that tries to propagate slide the compromises to adjacent edges. That process unfortunately is non-deterministic - it depends on the original ordering of vertices to some extent, though there is an effort to try to chain the propagation together after initial choices are made among connected components of edges. Unfortunately in the model here, different choices of propagation order are chosen for two of the edge chains vs the other two, resulting in different end results.

dhoni commented

2017-12-06 07:45:54 +01:00

Added subscriber: @dhoni