Regression: Bevel modifier spikes with the Loop Slide option #102532

New Issue

Adam Friesen · 2022-11-15T19:32:43+01:00

Adam Friesen commented

2022-11-15 19:32:43 +01:00

System Information
Operating system: Windows-10-10.0.19044-SP0 64 Bits
Graphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30

Blender Version
Broken: version: 3.5.0 Alpha, branch: master, commit date: 2022-11-15 00:55, hash: efcd587bc2
Worked: 2.93.11

NOTE: it still works without the Loop Slide option

The change in behavior is caused by 4886ed2863

Short description of error
Blender's beveling is heading back to being busted and broken, a simple subdivided surface with holes in it can no longer be beveled cleaning without explosions in the mesh.

Exact steps for others to reproduce the error
Open the .blend file and see for yourself, turn the bevel modifier off and then on again.

.blend
BrokenBevel.blend

**System Information** Operating system: Windows-10-10.0.19044-SP0 64 Bits Graphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30 **Blender Version** Broken: version: 3.5.0 Alpha, branch: master, commit date: 2022-11-15 00:55, hash: `efcd587bc2` Worked: 2.93.11 NOTE: it still works without the `Loop Slide` option The change in behavior is caused by 4886ed2863 **Short description of error** Blender's beveling is heading back to being busted and broken, a simple subdivided surface with holes in it can no longer be beveled cleaning without explosions in the mesh. **Exact steps for others to reproduce the error** Open the .blend file and see for yourself, turn the bevel modifier off and then on again. **.blend** [BrokenBevel.blend](https://archive.blender.org/developer/F13918697/BrokenBevel.blend)

Adam Friesen commented

2022-11-15 19:32:43 +01:00

Added subscriber: @Ace_Dragon

Germano Cavalcante commented

2022-11-16 22:28:32 +01:00

Added subscriber: @mano-wii

Germano Cavalcante commented

2022-11-16 22:28:32 +01:00

I can confirm that the behavior of the Bevel modifier changed in version 3.0.0 sometime between d08cc63e2f(May 4) and 84dcf12ceb(Sep 29).

Needs investigation.

I can confirm that the behavior of the Bevel modifier changed in version 3.0.0 sometime between `d08cc63e2f`(May 4) and `84dcf12ceb`(Sep 29). Needs investigation.

Philipp Oeser commented

2022-11-17 09:28:38 +01:00

Changed status from 'Needs Triage' to: 'Confirmed'

Philipp Oeser commented

2022-11-17 12:22:27 +01:00

Added subscribers: @howardt, @lichtwerk

Philipp Oeser commented

2022-11-17 12:22:27 +01:00

NOTE: it still works without the Loop Slide option

The change in behavior is caused by 4886ed2863
from the commit message:

The current bug showed that that angle was much too big, and limited
desired behavior in many cases. Changing the angle from 0.25 radians
to 0.0001 radians (about 0.006 degrees) still fixes the original bug
and seems very unlikely to be limiting desired behavior now.

CC @howardt : what to do now? Is there a better threshold possibly? Or expose it to the UI?

NOTE: it still works without the `Loop Slide` option The change in behavior is caused by 4886ed2863 from the commit message: > The current bug showed that that angle was much too big, and limited > desired behavior in many cases. Changing the angle from 0.25 radians > to 0.0001 radians (about 0.006 degrees) still fixes the original bug > and seems very unlikely to be limiting desired behavior now. CC @howardt : what to do now? Is there a better threshold possibly? Or expose it to the UI?

Philipp Oeser changed title from ~~Blender can no longer bevel a curved surface with holes without artifacts~~ to "Rgression": Bevel modifier spikes with the Loop Slide option

2022-11-17 12:23:42 +01:00

Pratik Borhade commented

2022-11-17 13:15:05 +01:00

Added subscriber: @PratikPB2123

Pratik Borhade commented

2022-11-17 13:15:05 +01:00

@lichtwerk hi, any improvements after this commit {d17f5bcd8f}? (Can not test ATM)

Philipp Oeser commented

2022-11-17 14:06:02 +01:00

In #102532#1447352, @PratikPB2123 wrote:
@lichtwerk hi, any improvements after this commit {d17f5bcd8f}? (Can not test ATM)

Not in this case, no

> In #102532#1447352, @PratikPB2123 wrote: > @lichtwerk hi, any improvements after this commit {d17f5bcd8f}? (Can not test ATM) Not in this case, no

Pratik Borhade changed title from ~~"Rgression": Bevel modifier spikes with the Loop Slide option~~ to Regression: Bevel modifier spikes with the Loop Slide option

2022-11-20 11:49:18 +01:00

Martijn Versteegh commented

2022-11-25 10:50:28 +01:00

Added subscriber: @Baardaap

Martijn Versteegh commented

2022-11-25 10:50:28 +01:00

It's just a single data point and I don't really know how to test this thoroughly. But trying various angles seemed to indicate this spiking (in this particular test file) stops with a BEVEL_GOOD_ANGLE of 0.015, and I think that still doesn't break #86768 though I'm not really sure what to test there as .25 in current master seemed to behave identical to 0.0001 and 0.015 for the test file provided, so probably I test the wrong thing.

Bruno commented

2022-12-26 23:41:17 +01:00

Added subscriber: @BlenderBruno

Bruno commented

2022-12-26 23:44:34 +01:00

FWIW, I confirm @Baardaap finding for the provided test file (using a custom build on 3.5 alpha). It starts to break down with a BEVEL_GOOD_ANGLE value lower than 0.0112655.

Untitled1.mp4

It seems the BEVEL_GOOD_ANGLE value can be as high as 0.16 without breaking #86768 and has no effect at all on the second test file of #86768 though (which still shows some unwanted artefacts with bevel modifier percent set to 100% with loopslide), attached here for reference:
bevel_loop_slide_failure2.blend

Untitled.mp4

FWIW, I confirm @Baardaap finding for the provided test file (using a custom build on 3.5 alpha). It starts to break down with a `BEVEL_GOOD_ANGLE` value lower than `0.0112655`. [Untitled1.mp4](https://archive.blender.org/developer/F14094074/Untitled1.mp4) It seems the `BEVEL_GOOD_ANGLE` value can be as high as `0.16` without breaking #86768 and has no effect at all on the second test file of #86768 though (which still shows some unwanted artefacts with bevel modifier percent set to 100% with loopslide), attached here for reference: [bevel_loop_slide_failure2.blend](https://archive.blender.org/developer/F9917825/bevel_loop_slide_failure2.blend) [Untitled.mp4](https://archive.blender.org/developer/F14094076/Untitled.mp4)

Thomas Dinges commented

2023-01-13 15:31:57 +01:00

Added subscriber: @ThomasDinges

Thomas Dinges commented

2023-01-13 15:31:57 +01:00

@howardt Hey Howard, any chance to take a look at this high priority report please? Thanks in advance.

Howard Trickey commented

2023-01-13 15:56:35 +01:00

Will look at it tomorrow.

Howard Trickey commented

2023-01-16 21:38:35 +01:00

I think I will change BEVEL_GOOD_ANGLE to 0.1, which works for all the test files in #44961, #86768, #95335, and #102532.
For some reason, I cannot get the modifiers test to work before making this change, so am afraid to make it for fear of breaking that test and not being able to fix it. I think I'll try on another computer.

I think I will change BEVEL_GOOD_ANGLE to 0.1, which works for all the test files in #44961, #86768, #95335, and #102532. For some reason, I cannot get the modifiers test to work before making this change, so am afraid to make it for fear of breaking that test and not being able to fix it. I think I'll try on another computer.

blender-admin commented

2023-01-22 16:28:12 +01:00

This issue was referenced by b544199c56

This issue was referenced by b544199c566bbb9529b5925ce96ae2d2abbf04c1

Howard Trickey commented

2023-01-22 16:44:16 +01:00

Changed status from 'Confirmed' to: 'Resolved'

Howard Trickey closed this issue

2023-01-22 16:44:16 +01:00

Howard Trickey self-assigned this 2023-01-22 16:44:16 +01:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Regression: Bevel modifier spikes with the Loop Slide option #102532