'Custom Split Normals' generated by the Boolean modifier does not match the normals of the object #95915

New Issue

tempdevnova · 2022-02-20T18:36:47+01:00

tempdevnova commented

2022-02-20 18:36:47 +01:00

System Information
Operating system: Windows-10-10.0.19041-SP0 64 Bits
Graphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.07

Blender Version
Broken: version: 3.0.0, branch: master, commit date: 2021-12-02 18:35, hash: f1cca30557

Short description of error
If the object used as operand in the Boolean modifier has Custom Split Normals (either through the WeightedNormal modifier or manually added) the resulting object will have a "broken" Custom Split Normal.
This kind of works with objects which do not have any Custom Split Normals (so only shade flat and shade smooth, no weighted normals etc.) (although it also has it's problems. See #95913 (Strange shading behaviour when using boolean modifier))
When Custom Split Normals are involved, the modifier uses the Custom Split Normals of the operand object on the target object to determine the shading on the affected areas.

Exact steps for others to reproduce the error

Open attached file. (The left example has only shade smooth on the boolean object and on the right example the boolean object has a weighted normal modifier on the modifier stack).
The issue still persists even if you delete the weighted normals modifier on the target object in the right example.
Test_file2.blend

Another example:
Test_file.blend

**System Information** Operating system: Windows-10-10.0.19041-SP0 64 Bits Graphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.07 **Blender Version** Broken: version: 3.0.0, branch: master, commit date: 2021-12-02 18:35, hash: `f1cca30557` **Short description of error** If the object used as operand in the Boolean modifier has `Custom Split Normals` (either through the `WeightedNormal` modifier or manually added) the resulting object will have a "broken" `Custom Split Normal`. This kind of works with objects which do not have any `Custom Split Normals` (so only shade flat and shade smooth, no weighted normals etc.) (although it also has it's problems. See #95913 (Strange shading behaviour when using boolean modifier)) When `Custom Split Normals` are involved, the modifier uses the `Custom Split Normals` of the operand object on the target object to determine the shading on the affected areas. **Exact steps for others to reproduce the error** - Open attached file. (The left example has only shade smooth on the boolean object and on the right example the boolean object has a weighted normal modifier on the modifier stack). The issue still persists even if you delete the weighted normals modifier on the target object in the right example. [Test_file2.blend](https://archive.blender.org/developer/F12878965/Test_file2.blend) --- Another example: [Test_file.blend](https://archive.blender.org/developer/F12878951/Test_file.blend)

tempdevnova commented

2022-02-20 18:36:47 +01:00

Added subscriber: @tempdevnova

Richard Antalik commented

2022-02-20 18:42:52 +01:00

Added subscriber: @iss

Richard Antalik commented

2022-02-20 18:42:52 +01:00

Wouldn't #95913 be the root cause here too?

tempdevnova commented

2022-02-20 19:23:20 +01:00

Well if you completely remove the normal transfering behaviour then i assume that problem would probably be fixed as well. But in case it was an intentional feature, this would have to be fixed seperately.

tempdevnova commented

2022-02-20 19:25:32 +01:00

Of course the blender devs could also get rid of this problematic behaviour even if it was intentional at first.

Germano Cavalcante commented

2022-02-22 01:37:38 +01:00

Added subscriber: @mano-wii

Germano Cavalcante commented

2022-02-22 01:37:38 +01:00

The first statement of fact can be confirmed and does not seem to be the desired result.
But I don't think the statement that "it simply doesn't work if any normal alterations are made after the boolean modifier" is correct.
In the attached files, although the result looks wrong, this is the expected result for the Weighted Normal with the Custom Split Normals generated by the Boolean modifier.
If you generate similar Custom Split Normals without the Boolean modifier the result is very similar.

Anyway, even if it wasn't the expected result of the Weighted Normal, it has nothing to do with Boolean and should be reported separately.

So, (if it's ok for you), I'll edit the description to report only the problem of the Custom Split Normals generated not matching the normals of the object.

The first statement of fact can be confirmed and does not seem to be the desired result. But I don't think the statement that `"it simply doesn't work if any normal alterations are made after the boolean modifier"` is correct. In the attached files, although the result looks wrong, this is the expected result for the `Weighted Normal` with the `Custom Split Normals` generated by the `Boolean` modifier. If you generate similar `Custom Split Normals` without the `Boolean` modifier the result is very similar. Anyway, even if it wasn't the expected result of the `Weighted Normal`, it has nothing to do with `Boolean` and should be reported separately. So, (if it's ok for you), I'll edit the description to report only the problem of the `Custom Split Normals` generated not matching the normals of the object.

Germano Cavalcante changed title from ~~Custom split normals break shading~~ to 'Custom Split Normals' generated by the Boolean modifier does not match the normals of the object

2022-02-22 01:43:12 +01:00

Germano Cavalcante commented

2022-02-22 01:43:39 +01:00

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

tempdevnova commented

2022-02-22 18:31:58 +01:00

"If you generate similar Custom Split Normals without the Boolean modifier the result is very similar." I could not reproduce similar results without the boolean modifier and I assume you actually meant the weighted normal modifier? If this is not the case could you please elaborate further.

But apart from that I also agree that the problem stems from the way the boolean modifier generates custom split normals.

Regarding that I strongly suggest a complete removal of the custom split normals generation by the boolean modifier as also mentioned in #95913.

"If you generate similar Custom Split Normals without the Boolean modifier the result is very similar." I could not reproduce similar results without the boolean modifier and I assume you actually meant the weighted normal modifier? If this is not the case could you please elaborate further. But apart from that I also agree that the problem stems from the way the boolean modifier generates custom split normals. Regarding that I strongly suggest a complete removal of the custom split normals generation by the boolean modifier as also mentioned in #95913.

Germano Cavalcante commented

2022-02-22 19:31:29 +01:00

In #95915#1312085, @tempdevnova wrote:
(...) could you please elaborate further. (...)

To create custom normals just click on the Add Custom Split Normals Data option in the Geometry Data panel (https://docs.blender.org/manual/en/dev/modeling/meshes/properties/custom_data.html#bpy-ops-mesh-customdata-custom-splitnormals-add)
One way to create similar custom normals to that of the Boolean is to increase the Auto Smooth angle to 180º and press the button.

> In #95915#1312085, @tempdevnova wrote: > (...) could you please elaborate further. (...) To create custom normals just click on the `Add Custom Split Normals Data` option in the `Geometry Data` panel (https://docs.blender.org/manual/en/dev/modeling/meshes/properties/custom_data.html#bpy-ops-mesh-customdata-custom-splitnormals-add) One way to create similar custom normals to that of the `Boolean` is to increase the `Auto Smooth` angle to 180º and press the button.

tempdevnova commented

2022-02-23 17:12:44 +01:00

"One way to create similar custom normals to that of the Boolean is to increase the Auto Smooth angle to 180º and press the button." Why is that exactly and do the shading errors also appear in these cases?

"Anyway, even if it wasn't the expected result of the Weighted Normal, it has nothing to do with Boolean and should be reported separately." I don't quite get what you mean here, do you think that the problem is caused by the weighted normals modifier or the boolean modifier?

"One way to create similar custom normals to that of the Boolean is to increase the Auto Smooth angle to 180º and press the button." Why is that exactly and do the shading errors also appear in these cases? "Anyway, even if it wasn't the expected result of the Weighted Normal, it has nothing to do with Boolean and should be reported separately." I don't quite get what you mean here, do you think that the problem is caused by the weighted normals modifier or the boolean modifier?

Germano Cavalcante commented

2022-02-23 17:28:20 +01:00

In my opinion, the result we see with the Weighted Normal is not a bug.
Custom Split Normals are normals that can be captured and edited by the user. If the object has Custom Split Normals, each Loop has a "custom" normal. The Weighted Normal modify these normals resulting in another one that can be weird in some extreme cases.

Also the submission template and guidelines mentions "If there are multiple bugs, make multiple bug reports" (although I'm almost certain that the result in Weighted Normal is not a bug).

In my opinion, the result we see with the `Weighted Normal` is not a bug. `Custom Split Normals` are normals that can be captured and edited by the user. If the object has `Custom Split Normals`, each Loop has a "custom" normal. The `Weighted Normal` modify these normals resulting in another one that can be weird in some extreme cases. Also the [submission template and guidelines](https://developer.blender.org/maniphest/task/edit/form/1/) mentions `"If there are multiple bugs, make multiple bug reports"` (although I'm almost certain that the result in `Weighted Normal` is not a bug).

tempdevnova commented

2022-02-23 18:51:27 +01:00

"The Weighted Normal modify these normals resulting in another one that can be weird in some extreme cases." Doesn't the weighted normal usually overwrite existing custom normals
"although I'm almost certain that the result in Weighted Normal is not a bug", yes I also think that this specific problem stems from the boolean modifier not the weighted normals, although the fact that the weighted normals doesn't overwrite the exisiting custom normals is still strange to me.

"If there are multiple bugs, make multiple bug reports" Yes I already tried to do that, I just wasn't sure whether the problem stemmed from the weighted normals or boolean modifier.
By the way how do you do the highlighting for your quotes and technical terms?

"The Weighted Normal modify these normals resulting in another one that can be weird in some extreme cases." Doesn't the weighted normal usually overwrite existing custom normals "although I'm almost certain that the result in Weighted Normal is not a bug", yes I also think that this specific problem stems from the boolean modifier not the weighted normals, although the fact that the weighted normals doesn't overwrite the exisiting custom normals is still strange to me. "If there are multiple bugs, make multiple bug reports" Yes I already tried to do that, I just wasn't sure whether the problem stemmed from the weighted normals or boolean modifier. By the way how do you do the highlighting for your quotes and technical terms?

Philipp Oeser removed the

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

'Custom Split Normals' generated by the Boolean modifier does not match the normals of the object #95915