Separate Transform node breaks when scale is negative. #147
Reference in New Issue
Block a user
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Original bug report: https://projects.blender.org/blender/blender/issues/157246
System Information
Operating system: windows
Graphics card: rtx2070
Blender Version
Broken: 5.2 release
Worked: unknown
separate transform does not work as expected when the scale is negative.
use combine transform to make a transform. set one or three of the axes to a negative value. plug into separate transform and back into another combine transform. or use transform geometry to view on something. when there are an odd number of negative scales the transform seems to be wrong.
Based on the default startup or an attached .blend file (as simple as possible).
Report is clear
LLM Reasoning
Got it, let's see. The bug report has a title that clearly states the issue: "Separate Transform node breaks when scale is negative." Then the body explains the problem: separate transform doesn't work as expected with negative scales. It gives steps to reproduce: use combine transform, set one or three axes to negative, plug into separate and back to combine, or use transform geometry. The description of the issue (odd number of negative scales causing wrong transform) is clear. So the bug report clearly describes the issue. So the answer is Yes.
JSON Data
{
"llm_judgement": true,
"processing_time": 6.9603459999999995
}
Report is specific
LLM Reasoning
Got it, let's analyze this bug report. The title is "Separate Transform node breaks when scale is negative." The body describes a specific scenario: using separate transform with negative scale values, mentioning steps like combining transforms, setting axes to negative, plugging into separate and back, or using transform geometry. It's about a particular issue (negative scale causing the node to break) rather than a vague problem. So it's a specific bug report.
JSON Data
{
"llm_judgement": true,
"processing_time": 6.312266999999999
}
Report flagged as duplicate by LLM
Separate transform node returns incorrect scale and rotation when scale is negative
https://projects.blender.org/blender/blender/issues/140589
System Information
Operating system: Windows 10
Graphics card: GTX 3060
Blender Version
Broken: (example: 4.2.9LTS)
Short description of error
When "Instance Transform" is fed directly into "Set Instance Transform", the result is identical, as expected.
However, if instance scale is negative, and "Instance Transform" is separated and combined (without other modification), the result is not as expected.
If combining transform from separate "Position", "Instance Rotation" and "Instance Scale" nodes, the result is once again correct.
It seems in fact that "Separate transform" node returns a positive scale even when instance scale is negative.
Exact steps for others to reproduce the error
Joined files incorporate a geometry nodes which can switch between three use cases.
See how "RecombinedTransform" gives a different result than other two modes when using negative scales.
LLM Reasoning
Got it, let's analyze these two bug reports. First, look at the descriptions and steps to reproduce.
Bug report 1 says: Separate Transform node breaks when scale is negative. Steps involve using combine transform with negative axes, plugging into separate, then back into combine or using transform geometry. The issue is that with odd number of negative scales, the transform is wrong.
Bug report 2 says: Separate transform returns incorrect scale and rotation when scale is negative. Specifically, it mentions that when instance scale is negative, separating and combining gives a different result than expected. Also, "Separate transform" returns positive scale even if input was negative.
Both are about the Separate Transform node not handling negative scales correctly. The core issue seems to be the same: negative scale in separate transform leads to incorrect output (either wrong transform or positive scale instead of negative). So they're duplicates.
JSON Data
{
"llm_judgement": true,
"processing_time": 10.885824999999999,
"duplicate_report_number": "140589"
}
Report flagged as NOT a duplicate by LLM
"Transform" node distorts when scale is not 1
https://projects.blender.org/blender/blender/issues/94820
System Information
Operating system: Windows-10-10.0.19043-SP0 64 Bits
Graphics card: NVIDIA GeForce GTX 760/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11
Blender Version
Broken: version: 3.0.0, branch: master, commit date: 2021-12-02 18:35, hash:
f1cca30557Worked: N/A
Short description of error
When a mesh has a scale other than (1, 1, 1), the "Transform" node distorts the mesh in a strange way.
It can be fixed with "Apply Scale", but it shouldn't be necessary to apply the scale.
Exact steps for others to reproduce the error
The attached .blend file demonstrates the distortion.
strangetransform.blend
odd.mp4
LLM Reasoning
Got it, let's analyze the two bug reports. First, look at the descriptions and steps to reproduce.
Bug report 1 is about the "Separate Transform" node breaking when scale is negative. The steps involve using Combine Transform, setting negative scales on axes, plugging into Separate Transform, then back into another Combine, or using Transform Geometry. The issue is with odd number of negative scales causing wrong transform.
Bug report 2 is about the "Transform" node distorting when scale isn't (1,1,1). Steps: take a mesh, scale it (e.g., X by 10), apply Transform node and rotate Z, notice distortion. The problem here is scaling the mesh before using the Transform node leading to distortion.
The key difference: Bug 1 is about Separate Transform with negative scales, Bug 2 is about Transform node with non-1 scales causing distortion. The issues are different—one is about negative scale in Separate Transform, the other is about any non-1 scale (positive or negative?) in Transform node. So they aren't the same issue. Therefore, the answer is No.
JSON Data
{
"llm_judgement": false,
"processing_time": 14.554396,
"duplicate_report_number": "94820"
}
Report flagged as NOT a duplicate by LLM
Wrong negative separated units
https://projects.blender.org/blender/blender/issues/21381
%%%This is using Blender 2.5 Alpha 1.
I changed the displayed units in Scene -> Units, and turned on the Separate Units option.
A value of -1.227m became -2m 77.3cm which is clearly wrong.
This only affected the UI as far as I can tell.%%%
LLM Reasoning
Got it, let's analyze these two bug reports. First, look at the descriptions and steps to reproduce.
Bug report 1 is about the Separate Transform node in Blender (version 5.2) where negative scale causes issues—like when combining transforms with negative scales, the separate transform doesn't work as expected, especially with odd numbers of negative axes. The steps involve using combine/transform nodes.
Bug report 2 is about unit conversion in Blender 2.5 Alpha 1: a value -1.227m becomes incorrectly displayed as -2m 77.3cm when Separate Units is on. This is about UI display of units, not transform nodes or scale issues.
The key sections are "Description of issue" and "Steps to reproduce". Report 1 is about transform node behavior with negative scales; report 2 is about unit conversion display. They're completely different topics—transform vs. units. So they aren't the same issue. Therefore, the answer is No.
JSON Data
{
"llm_judgement": false,
"processing_time": 11.665876,
"duplicate_report_number": "21381"
}
Report flagged as NOT a duplicate by LLM
Transform constraint maps wrongly with negative scale
https://projects.blender.org/blender/blender/issues/27886
%%%svn 38162M
Open the attached file.
There are two bones.
Y scale of the green bone is controlled through a Transform constraint with the grey bone as source.
Mapping is as follows:
Source range -1 to 1 maps to destination 0 to 2
so theoretically, scale should be mapped like this:
(grey bone source -> green bone destination)
0 -> 1
1 -> 2
actually, it maps like this
0 -> 1
1 -> 2
So for the negative values, the absolute value is calculated before mapping, instead of a linear transformation as would be expected. It is not possible to use negative scaling of the source at all.
same problem occurs for all axes.
%%%
LLM Reasoning
Got it, let's analyze these two bug reports. First, look at the descriptions and steps.
Bug report 1 talks about "Separate Transform node" in Blender (probably a node setup) where negative scale causes issues when using Combine/Transform nodes. The problem is that with odd number of negative scales, the transform is wrong.
Bug report 2 is about a "Transform constraint" where negative scale on the source bone maps incorrectly—negative values get absolute value before mapping instead of linear transformation.
So the first is about node-based transforms (Separate/Combine), the second is about bone constraints. Different contexts: one is nodes, the other is armature constraints. The core issue might be similar (negative scale handling), but the components are different (nodes vs constraints). Since they're in different parts of Blender (node system vs constraint system), they're not the same issue. So the answer is No.
JSON Data
{
"llm_judgement": false,
"processing_time": 11.136807999999998,
"duplicate_report_number": "27886"
}
Report flagged as duplicate by LLM
FunctionNodeSeparateTransform nodes does not support negative scale
https://projects.blender.org/blender/blender/issues/138967
System Information
Operating system: Windows-10-10.0.19041-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 4070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 551.31 Opengl Backend
Blender Version
Broken: version: 4.5.0 Alpha, branch: main (modified), commit date: 2025-05-16 05:59, hash:
1232d64681b7Worked: (newest version of Blender that worked as expected)
Short description of error
[Please fill out a short description of the error here]
Exact steps for others to reproduce the error
[Please describe the exact steps needed to reproduce the issue]
[Based on the default startup or an attached .blend file (as simple as possible)]
open the blend file below,the result of the gn is wrong
LLM Reasoning
Got it, let's analyze these two bug reports. First, look at the titles: both mention issues with Separate Transform nodes and negative scale. The first report says "Separate Transform node breaks when scale is negative" and describes steps where using combine transform with negative scales leads to wrong transforms, especially odd numbers of negatives. The second report's title is "FunctionNodeSeparateTransform nodes does not support negative scale" and mentions the result is wrong. Both are about Separate Transform nodes failing with negative scale values. The descriptions focus on the same core issue: negative scale causing problems in the separate transform node. So they seem to be duplicates.
JSON Data
{
"llm_judgement": true,
"processing_time": 8.669725999999999,
"duplicate_report_number": "138967"
}