Sample Volume node's Position input is off by half the voxel width #113265

Open
opened 2023-10-04 23:20:36 +02:00 by Cornus-Ammonis · 4 comments

System Information
Operating system: Windows-10-10.0.19045-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 536.23

Blender Version
Broken: version: 4.1.0 Alpha, branch: main, commit date: 2023-09-29 20:55, hash: 90ec11b823e8

Short description of error
The Sample Volume node's Position input is off by half the voxel width.

Exact steps for others to reproduce the error
This issue is easiest to notice when used with Simulation Nodes because the error accumulates over time. In the attached blend file, a checkbox controls whether to apply a fix to the error or not. If the fix is not applied, the mesh will drift in the -X, -Y, -Z direction over time. When working correctly, the mesh does not drift. In order to fix it, half the voxel width is subtracted from the Sample Volume node's position input.

**System Information** Operating system: Windows-10-10.0.19045-SP0 64 Bits Graphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 536.23 **Blender Version** Broken: version: 4.1.0 Alpha, branch: main, commit date: 2023-09-29 20:55, hash: `90ec11b823e8` **Short description of error** The Sample Volume node's Position input is off by half the voxel width. **Exact steps for others to reproduce the error** This issue is easiest to notice when used with Simulation Nodes because the error accumulates over time. In the attached blend file, a checkbox controls whether to apply a fix to the error or not. If the fix is not applied, the mesh will drift in the -X, -Y, -Z direction over time. When working correctly, the mesh does not drift. In order to fix it, half the voxel width is subtracted from the Sample Volume node's position input.
Cornus-Ammonis added the
Severity
Normal
Type
Report
Status
Needs Triage
labels 2023-10-04 23:20:37 +02:00

Confirmed and I've attached a slightly simplified/more visual example .blend showing the offset. This is similar to #85312 in that OpenVDB has a different idea of what the center of a voxel is compared to blender. This seems like a similar situation almost.

Confirmed and I've attached a slightly simplified/more visual example .blend showing the offset. This is similar to #85312 in that OpenVDB has a different idea of what the center of a voxel is compared to blender. This seems like a similar situation almost.
Jesse Yurkovich added
Module
Nodes & Physics
Status
Confirmed
and removed
Status
Needs Triage
labels 2023-10-05 01:05:56 +02:00
@LukasTonne /
Iliya Katushenock added the
Interest
Geometry Nodes
label 2023-10-05 17:21:28 +02:00
Member

I don't think the point sampling is actually wrong. It uses the position inputs as they are, so a position (0, 0, 0) would sample right at voxel (0, 0, 0), assuming there is no translation of the grid. OpenVDB grids don't care which point inside a voxel the coordinate represents, but generally we'd consider this to be the voxel center.

Some of the volume nodes offset the grids by half a voxel to compensate for this:

Crucially, the Mesh to Volume, Mesh to SDF Volume, and SDF Volume Sphere node do not add any translation. I believe this is where the "drift" in the original file comes from: the mesh is converted back and forth to a volume but the Volume to Mesh node adds an offset that is not reverted by the Mesh to Volume node.

I don't think the point sampling is actually wrong. It uses the position inputs as they are, so a position (0, 0, 0) would sample right at voxel (0, 0, 0), assuming there is no translation of the grid. OpenVDB grids don't care which point inside a voxel the coordinate represents, but generally we'd consider this to be the voxel center. Some of the volume nodes offset the grids by half a voxel to compensate for this: - [Volume Cube](https://projects.blender.org/blender/blender/src/commit/9a654c04dea839ef8da6ef006abc54a90fff86d3/source/blender/nodes/geometry/nodes/node_geo_volume_cube.cc#L168) changes the grid transform - [Volume to Mesh](https://projects.blender.org/blender/blender/src/commit/9a654c04dea839ef8da6ef006abc54a90fff86d3/source/blender/blenkernel/intern/volume_to_mesh.cc#L106) modifies the mesh vertex positions afterwards Crucially, the _Mesh to Volume_, _Mesh to SDF Volume_, and _SDF Volume Sphere_ node do **not** add any translation. I believe this is where the "drift" in the original file comes from: the mesh is converted back and forth to a volume but the _Volume to Mesh_ node adds an offset that is not reverted by the _Mesh to Volume_ node.
Author

Wouldn't you expect a loop of only Mesh to Volume -> Volume to Mesh to exhibit the same drift if that's the case? That does not occur in my testing.

Wouldn't you expect a loop of only Mesh to Volume -> Volume to Mesh to exhibit the same drift if that's the case? That does not occur in my testing.
Bart van der Braak added
Type
Bug
and removed
Type
Report
labels 2024-08-14 13:07:55 +02:00
Sign in to join this conversation.
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset System
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Code Documentation
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
Interest
Freestyle
Interest
Geometry Nodes
Interest
Grease Pencil
Interest
ID Management
Interest
Images & Movies
Interest
Import Export
Interest
Line Art
Interest
Masking
Interest
Metal
Interest
Modeling
Interest
Modifiers
Interest
Motion Tracking
Interest
Nodes & Physics
Interest
OpenGL
Interest
Overlay
Interest
Overrides
Interest
Performance
Interest
Physics
Interest
Pipeline, Assets & IO
Interest
Platforms, Builds & Tests
Interest
Python API
Interest
Render & Cycles
Interest
Render Pipeline
Interest
Sculpt, Paint & Texture
Interest
Text Editor
Interest
Translations
Interest
Triaging
Interest
Undo
Interest
USD
Interest
User Interface
Interest
UV Editing
Interest
VFX & Video
Interest
Video Sequencer
Interest
Viewport & EEVEE
Interest
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
Legacy
Asset Browser Project
Legacy
Blender 2.8 Project
Legacy
Milestone 1: Basic, Local Asset Browser
Legacy
OpenGL Error
Meta
Good First Issue
Meta
Papercut
Meta
Retrospective
Meta
Security
Module
Animation & Rigging
Module
Core
Module
Development Management
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline, Assets & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Module
Viewport & EEVEE
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Severity
High
Severity
Low
Severity
Normal
Severity
Unbreak Now!
Status
Archived
Status
Confirmed
Status
Duplicate
Status
Needs Info from Developers
Status
Needs Information from User
Status
Needs Triage
Status
Resolved
Type
Bug
Type
Design
Type
Known Issue
Type
Patch
Type
Report
Type
To Do
No Milestone
No project
No Assignees
4 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#113265
No description provided.