Sculpt: Clay strip sculpts on back face when front face only is turned on #116458

New Issue

HF · 2023-12-22T02:25:44+01:00

HF commented

2023-12-22 02:25:44 +01:00

System Information
Operating system: Windows-10-10.0.19042-SP0 64 Bits
Graphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 546.33

Blender Version
Broken: version: 4.0.1, branch: blender-v4.0-release, commit date: 2023-11-16 16:40, hash: d0dd92834a08
Worked: (newest version of Blender that worked as expected)

Short description of error
sculpt on back face (or occluded region) when front face only is turned on see video

Exact steps for others to reproduce the error
Turn on sculpt front face only, make back face really close to front face, sculpt.

bug.blend
video1
video2

**System Information** Operating system: Windows-10-10.0.19042-SP0 64 Bits Graphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 546.33 **Blender Version** Broken: version: 4.0.1, branch: blender-v4.0-release, commit date: 2023-11-16 16:40, hash: `d0dd92834a08` Worked: (newest version of Blender that worked as expected) **Short description of error** sculpt on back face (or occluded region) when front face only is turned on see video **Exact steps for others to reproduce the error** Turn on sculpt front face only, make back face really close to front face, sculpt. [bug.blend](https://projects.blender.org/attachments/81fcaf32-e049-438a-b094-f8720d5a7eba) [video1](https://projects.blender.org/attachments/6d402293-4d63-4e46-a36c-995126592a1a) [video2](https://projects.blender.org/attachments/1670dfb8-3bab-44ba-b8e5-1cd2398f288a)

blender_pN8ftkAZ3Q.mp4

3.0 MiB

HF added the

labels 2023-12-22 02:25:45 +01:00

Pratik Borhade commented

2023-12-22 05:32:10 +01:00

Hi, thanks for the report. I'm not able to confirm this in 4.1
Could it be that the normals are pointing outwards (i.e. front face). Have you checked with face orientation overlay?
Can you share .blend file?

Hi, thanks for the report. I'm not able to confirm this in 4.1 Could it be that the normals are pointing outwards (i.e. front face). Have you checked with `face orientation` overlay? Can you share .blend file?

Pratik Borhade added

Status

Needs Information from User

and removed

Status

Needs Triage

labels 2023-12-22 05:32:20 +01:00

HF commented

2023-12-22 06:47:14 +01:00

orientation is fine

here is another video, this time it sculpted on the other side of the mouth, I thought topology auto masking is supposed to prevent that, and based on how the other side is behaving, it basically broke, I think it's a bug

![image](/attachments/16db2f65-51b2-40c0-806f-94a56a4e9fab) orientation is fine here is another video, this time it sculpted on the other side of the mouth, I thought topology auto masking is supposed to prevent that, and based on how the other side is behaving, it basically broke, I think it's a bug

image.png

645 KiB

blender_yGCwYI6DJW.mp4

13 MiB

HF commented

2023-12-22 07:00:32 +01:00

this only happens when the brush is too big im using the clay strip brush

bug.blend

20 MiB

blender_Qsb52ZYFnJ.mp4

11 MiB

Pratik Borhade commented

2023-12-22 07:01:57 +01:00

Thanks. I see. So you are sculpting on front faces but somehow other faces of the mesh are affected by strokes. Could you check if any global auto-masking option is causing this: image

Thanks. I see. So you are sculpting on front faces but somehow other faces of the mesh are affected by strokes. Could you check if any global auto-masking option is causing this: [image](/attachments/09ee7906-cb97-4a01-9cc6-5a5a689228eb)

image.png

52 KiB

HF commented

2023-12-22 07:12:58 +01:00

no, nothing is checked there

Pratik Borhade commented

2023-12-22 08:01:32 +01:00

Thanks, can confirm in your file. Possibly due to small object size and falloff. Though this does not happen with other brushes.

Pratik Borhade added

Module

Sculpt, Paint & Texture

Status

Confirmed

and removed

Status

Needs Information from User

labels 2023-12-22 08:01:43 +01:00

Pratik Borhade changed title from ~~sculpt on back face when front face only is turned on~~ to Sculpt: Clay strip scllupts on back face when front face only is turned on

2023-12-22 08:02:12 +01:00

Pratik Borhade changed title from ~~Sculpt: Clay strip scllupts on back face when front face only is turned on~~ to Sculpt: Clay strip sculpts on back face when front face only is turned on

2023-12-22 08:02:26 +01:00

Prikshit singh referenced a pull request that will close this issue

2023-12-23 17:25:05 +01:00

Fix #116458: Clay strip sculpts on back face when front face only is turned on #116493

Julien Kaspar commented

2024-01-04 10:25:56 +01:00

I don't think this has anything to do with the "Front Faces Only" option. That option just excludes faces that don't face the view.

This bug has been the most crippling issue for using any of these brushes imo: Clay, Clay Strips, Flatten, Fill & Scrape.
I think the issue became more noticable roughly around 2.81 when the area plane position was shifted to be at the exact cursor position instead of the mean point of the geometry within the cursor.

The issue seems to be that the falloff curve is not used for the flattening effect these brushes have. The full Strength is always used for flattening, resulting in geometry in the outer radius of the brush being pulled just as strong as in the center.

I have a couple of ideas:

An easy fix could be to use the falloff curve for the flattening effect of the brushes as well.
Although this will not work for brushes like Clay Strips that use the "Hardness" and "Tip Roundness" settings to shift the 0/0 point of the falloff curve.
Testing it would be good though.

It could also help to introduce falloff settings to scale the height and depth of the falloff along the Z axis of the area plane. Basically like the "Plane Trim" setting on Flatten brushes, but relative to the brush radius and with a falloff curve instead of as a bool.
That way the ripping of unwanted geometry can be reduced for Clay, Fill and Scrape brushes in the direction that suits them.

I don't think this has anything to do with the "Front Faces Only" option. That option just excludes faces that don't face the view. This bug has been the most crippling issue for using any of these brushes imo: Clay, Clay Strips, Flatten, Fill & Scrape. I think the issue became more noticable roughly around 2.81 when the area plane position was shifted to be at the exact cursor position instead of the mean point of the geometry within the cursor. ![image](/attachments/9959fa3f-e744-4e53-8d40-5ff0e9858538) The issue seems to be that the falloff curve is not used for the flattening effect these brushes have. The full Strength is always used for flattening, resulting in geometry in the outer radius of the brush being pulled just as strong as in the center. I have a couple of ideas: An easy fix could be to use the falloff curve for the flattening effect of the brushes as well. Although this will not work for brushes like Clay Strips that use the "Hardness" and "Tip Roundness" settings to shift the 0/0 point of the falloff curve. Testing it would be good though. It could also help to introduce falloff settings to scale the height and depth of the falloff along the Z axis of the area plane. Basically like the "Plane Trim" setting on Flatten brushes, but relative to the brush radius and with a falloff curve instead of as a bool. That way the ripping of unwanted geometry can be reduced for Clay, Fill and Scrape brushes in the direction that suits them.

image.png

305 KiB

Julien Kaspar commented

2024-01-08 14:59:48 +01:00

I am a bit confused how the falloff is used for brushes like Clay Strips.
It can easily seem like the Falloff isn't a "Sphere" like the brush settings imply, but instead a cylinder (sculpt plane with depth).
The Clay Strips falloff also seems to have far more depth than the radius.

Some more investigation in this issue would be nice.

I am a bit confused how the falloff is used for brushes like Clay Strips. It can easily seem like the Falloff isn't a "Sphere" like the brush settings imply, but instead a cylinder (sculpt plane with depth). The Clay Strips falloff also seems to have far more depth than the radius. Some more investigation in this issue would be nice.

Julien Kaspar added this to the Sculpt, Paint & Texture project 2024-02-13 17:56:13 +01:00

Raul Fernandez Hernandez self-assigned this 2024-02-20 05:08:24 +01:00

Raul Fernandez Hernandez commented

2024-02-20 05:35:28 +01:00

Is more obvious in those brushes but any brush with large enough radius relative to the mesh will start altering vertices in its sphere of influence messing up with that type of geometry with occluded parts.
Along with @JulienKaspar suggestions about the falloffs it may be worth exploring occlusion as an extra constraint.

Is more obvious in those brushes but any brush with large enough radius relative to the mesh will start altering vertices in its sphere of influence messing up with that type of geometry with occluded parts. Along with @JulienKaspar suggestions about the falloffs it may be worth exploring occlusion as an extra constraint.

Julien Kaspar commented

2024-02-20 11:00:23 +01:00

@farsthary I agree that any brush can have this behavior. But the flattening towards the sculpt plane is too strong/deep in particular.

Occlusion might not be a good fit here. It will be too aggressive in excluding vertices in surface details and cause a lot of noise.
Maybe a simple solution like adjusting the falloff shape for the flattening effect is enough to work?

Another idea could be to use geodesic surface distances based on the brush size instead of a sphere.
This was already prototyped once for the grab brush. Having this work in a performant way on any brush could solve the issue.
The idea back then was to use this behavior when "Topology Auto-Masking" is enabled (use_automasking_topology).
In fact, that sort of geodesic distance implementation could solve a lot of issues with accidentally affecting backfacing and disconnected geometry. Right now we mostly rely on exclusion based on normal direction which tends to fail.

@farsthary I agree that any brush can have this behavior. But the flattening towards the sculpt plane is too strong/deep in particular. Occlusion might not be a good fit here. It will be too aggressive in excluding vertices in surface details and cause a lot of noise. Maybe a simple solution like adjusting the falloff shape for the flattening effect is enough to work? Another idea could be to use geodesic surface distances based on the brush size instead of a sphere. This was [already prototyped once for the grab brush](https://archive.blender.org/developer/differential/0010/0010068/#251636). Having this work in a performant way on any brush could solve the issue. The idea back then was to use this behavior when "Topology Auto-Masking" is enabled (`use_automasking_topology`). In fact, that sort of geodesic distance implementation could solve a lot of issues with accidentally affecting backfacing and disconnected geometry. Right now we mostly rely on exclusion based on normal direction which tends to fail.

👍 2

Raul Fernandez Hernandez commented

2024-02-20 12:54:33 +01:00

Yes, geodesic distance will be ideal for this, may be better to introduce it in a follow up enhancement PR and for this one improving the falloffs.

👍 1

Raul Fernandez Hernandez referenced a pull request that will close this issue

2024-02-24 16:07:03 +01:00

WIP: Fix #116458: Added decay factor for flattening brushes. #118699

Sign in to join this conversation.