Principled shader clearcoat roughness inconsistent change from 0.0265 to 0.0266 (Cycles only) #114940

New Issue

Open

opened 2023-11-15 22:56:58 +01:00 by Tomás Cejas · 2 comments

Tomás Cejas commented

2023-11-15 22:56:58 +01:00

Windows-10-10.0.19041-SP0 64 Bits
NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 546.01

Broken: version: 4.0.0, branch: blender-v4.0-release, commit date: 2023-11-13 17:26, hash: 878f71061b8e.
Version 4.1 alpha.
3.X versions presented a similar issue (#53038).

Worked: unknown

Below a clearcoat roughness value of 0.0266, it seemingly drops to zero, impeding the making of very (but not perfectly) glossy clearcoats. In EEVEE this issue is not present, the clearcoat roughness smoothly transitions to zero. The issue is similar to #53038.

The following image ilustrates the problem, indicating clearcoat roughness values on a principled shader. Base roughness is set to 0 to better visualize the clearcoat reflections.

To visualize the problem, simply observe how the glossy reflection of a light changes as the clearcoat roughness value or a principled BSDF shader approaches 0. The file used to make the image is provided.

Windows-10-10.0.19041-SP0 64 Bits NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 546.01 Broken: version: 4.0.0, branch: blender-v4.0-release, commit date: 2023-11-13 17:26, hash: `878f71061b8e`. Version 4.1 alpha. 3.X versions presented a similar issue (https://projects.blender.org/blender/blender/issues/53038). Worked: unknown Below a clearcoat roughness value of 0.0266, it seemingly drops to zero, **impeding the making of very (but not perfectly) glossy clearcoats**. In EEVEE this issue is not present, the clearcoat roughness smoothly transitions to zero. The issue is similar to https://projects.blender.org/blender/blender/issues/53038. The following image ilustrates the problem, indicating clearcoat roughness values on a principled shader. Base roughness is set to 0 to better visualize the clearcoat reflections. ![image](/attachments/e4a13130-36d3-4603-8d1d-011af61b6b84) To visualize the problem, simply observe how the glossy reflection of a light changes as the clearcoat roughness value or a principled BSDF shader approaches 0. The file used to make the image is provided.

61 KiB

Clearcoat roughness.blend

774 KiB

Tomás Cejas added the

labels 2023-11-15 22:56:58 +01:00

Tomás Cejas referenced this issue

2023-11-15 22:59:44 +01:00

Principled shader clearcoat roughness inconsistent change from 0.0177 to 0.0178. #53038

Iliya Katushenock added the

Render & Cycles

label 2023-11-16 19:21:02 +01:00

Richard Antalik added

Render & Cycles

and removed

labels 2023-12-07 19:07:23 +01:00

Richard Antalik commented

2023-12-07 19:08:13 +01:00

Member

I think I have seen very similar recent report, but it could be concerning different property.

I think I have seen very similar recent report, but it could be concerning different property.

Weizhen Huang commented

2024-02-19 20:33:10 +01:00

Member

This is because we have a threshold where we decide the surface is effectively smooth. Currently the value is 5e-7, which corresponds to 0.02659 in the UI. Apparently this value is not small enough.
According to my experiment 1e-10 is a good value so that the switching is not obvious for bsdf_eval(), I can change that easily, however this value causes numerical issues for bsdf_sample(). The solution is to return eval / pdf instead of eval in bsdf_sample(). Seems that a lot of places need to be changed accordingly, but we should do that anyway. I will find some time to address this.
Also, this is not only a problem for coat, but all closures using microfacet distribution.

This is because we have a threshold where we decide the surface is effectively smooth. Currently the value is `5e-7`, which corresponds to `0.02659` in the UI. Apparently this value is not small enough. According to my experiment `1e-10` is a good value so that the switching is not obvious for `bsdf_eval()`, I can change that easily, however this value causes numerical issues for `bsdf_sample()`. The solution is to return `eval / pdf` instead of `eval` in `bsdf_sample()`. Seems that a lot of places need to be changed accordingly, but we should do that anyway. I will find some time to address this. Also, this is not only a problem for coat, but all closures using microfacet distribution.

Weizhen Huang self-assigned this 2024-02-19 20:33:21 +01:00

Sign in to join this conversation.

No Label

Animation & Rigging

Asset Browser Project Overview

Automated Testing

Blender Asset Bundle

Dependency Graph

Development Management

EEVEE & Viewport

Images & Movies

Motion Tracking

Nodes & Physics

Pipeline, Assets & IO

Platforms, Builds & Tests

Render & Cycles

Render Pipeline

Sculpt, Paint & Texture

Video Sequencer

Virtual Reality

Blender 2.8 Project

Milestone 1: Basic, Local Asset Browser

Good First Issue

Animation & Rigging

Development Management

EEVEE & Viewport

Nodes & Physics

Pipeline, Assets & IO

Platforms, Builds & Tests

Render & Cycles

Sculpt, Paint & Texture

Needs Info from Developers

Needs Information from User

No Milestone

No project

No Assignees

3 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#114940

No description provided.