Topology Mirror usage #81698

New Issue

Mahindra Ch · 2020-10-14T11:24:04+02:00

Mahindra Ch commented

2020-10-14 11:24:04 +02:00

System Information
Operating system: Windows-10-10.0.18362-SP0 64 Bits
Graphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38

Blender Version
Broken: version: 2.90.1, branch: master, commit date: 2020-09-23 06:43, hash: blender/blender@3e85bb34d0
Worked: (newest version of Blender that worked as expected)

Short description of error
[Vertex bug :: moving one vertex move other vertex which is not selected.{F8989425}]

Exact steps for others to reproduce the error
[I can't reproduce because it seems to happen randomly.]
Desktop 2020.10.14 - 14.56.01.01.mp4

**System Information** Operating system: Windows-10-10.0.18362-SP0 64 Bits Graphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38 **Blender Version** Broken: version: 2.90.1, branch: master, commit date: 2020-09-23 06:43, hash: `blender/blender@3e85bb34d0` Worked: (newest version of Blender that worked as expected) **Short description of error** [Vertex bug :: moving one vertex move other vertex which is not selected.{[F8989425](https://archive.blender.org/developer/F8989425/Bug.blend)}] **Exact steps for others to reproduce the error** [I can't reproduce because it seems to happen randomly.] [Desktop 2020.10.14 - 14.56.01.01.mp4](https://archive.blender.org/developer/F8989435/Desktop_2020.10.14_-_14.56.01.01.mp4)

Mahindra Ch commented

2020-10-14 11:24:04 +02:00

Added subscriber: @MOSESNES

blender-admin commented

2020-10-14 11:24:04 +02:00

blender/blender#101997 was marked as duplicate of this issue

blender-admin commented

2020-10-14 11:24:04 +02:00

blender/blender#100970 was marked as duplicate of this issue

blender-admin commented

2020-10-14 11:24:04 +02:00

blender/blender#91026 was marked as duplicate of this issue

Scott Ramsay commented

2020-10-14 12:17:04 +02:00

Added subscriber: @smramsay

Scott Ramsay commented

2020-10-14 12:17:04 +02:00

The bug appears to be caused by the "Topology Mirror" option in the 3D Viewport's "Options" popover.
Disabling that stops the problem from happening.
The other vertex is snapping to the X-mirrored location of the vertex being moved.
The vertices have consecutive indices if that hints at anything as well.

Bug appears in LTS 2.83.7

The bug appears to be caused by the "Topology Mirror" option in the 3D Viewport's "Options" popover. Disabling that stops the problem from happening. The other vertex is snapping to the X-mirrored location of the vertex being moved. The vertices have consecutive indices if that hints at anything as well. Bug appears in LTS 2.83.7

Philipp Oeser commented

2020-10-14 12:51:47 +02:00

Added subscribers: @Blendify, @ideasman42, @lichtwerk

Philipp Oeser commented

2020-10-14 12:51:47 +02:00

Changed status from 'Needs Triage' to: 'Needs Developer To Reproduce'

Philipp Oeser commented

2020-10-14 12:51:47 +02:00

Hm, Topology Mirror is somewhat awkward (and I think there might be room for improvement -- or the documentation has).

https://docs.blender.org/manual/en/dev/modeling/meshes/tools/tool_settings.html#topology-mirror

Topology Mirror tries to address this problem by determining which vertices are mirrored vertices not only by using their positions but also by looking at how those vertices are related to others in the Mesh Geometry. It looks at the overall topology to determine if particular vertices will be treated as mirrored. The effect of this is that mirrored vertices can be non-symmetrical and yet still be treated as mirrored.

Here is another explanation:

Blender uses internally an ID number for each vertex, which is calculated for the purpose of this tool. This special number is based on the surrounding topology of each vertex, taking into acount surrounding vertices/edges count (not its 3D position, of course). If the vertex matches in topology with the one on the other side, they will both have the same ID number, whereas vertices with different surrounding topology will end up having different IDs.
So in a model, every pair of matching vertices will have the same ID; we will say that this ID will match. And very important, too, that ID will be unique for that pair, provided that no other pais has the very same surrounding topology, which is difficult in a complex model.

Pretty sure this will not be considered a bug, but what I find confusing is that this example file clearly shows that picking candidates for mirroring with Topology Mirror does not really care about the chosen axis at all, so picking candidates is solely based on topology, not 3D positions. This might be intended [it might also be confusing that a candidate can be on the same side of chosen axis], but then the documentation could at least explain this a bit.

Reaching out to @ideasman42 or @Blendify

Hm, `Topology Mirror` is somewhat awkward (and I think there might be room for improvement -- or the documentation has). https://docs.blender.org/manual/en/dev/modeling/meshes/tools/tool_settings.html#topology-mirror > Topology Mirror tries to address this problem by determining which vertices are mirrored vertices not only by using their positions but also by looking at how those vertices are related to others in the Mesh Geometry. It looks at the overall topology to determine if particular vertices will be treated as mirrored. The effect of this is that mirrored vertices can be non-symmetrical and yet still be treated as mirrored. Here is another explanation: > Blender uses internally an ID number for each vertex, which is calculated for the purpose of this tool. This special number is based on the surrounding topology of each vertex, taking into acount surrounding vertices/edges count (not its 3D position, of course). If the vertex matches in topology with the one on the other side, they will both have the same ID number, whereas vertices with different surrounding topology will end up having different IDs. > So in a model, every pair of matching vertices will have the same ID; we will say that this ID will match. And very important, too, that ID will be unique for that pair, provided that no other pais has the very same surrounding topology, which is difficult in a complex model. Pretty sure this will not be considered a bug, but what I find confusing is that this example file clearly shows that picking candidates for mirroring with Topology Mirror does not really care about the chosen axis at all, so picking candidates is solely based on topology, not 3D positions. This might be intended [it might also be confusing that a candidate can be on the same side of chosen axis], but then the documentation could at least explain this a bit. Reaching out to @ideasman42 or @Blendify

Philipp Oeser changed title from ~~Vertex bug~~ to Topology Mirror usage

2020-10-14 12:52:40 +02:00

Roman commented

2020-11-01 12:32:33 +01:00

Added subscriber: @rwman