Adding very weak emissions to the scene weakens the highlights in cycles render #73328

Closed
opened 2020-01-22 23:03:10 +01:00 by Cubimation · 13 comments

System Information
Operating system:macOS 10.15.2
Graphics card:Radeon Pro 580X 8 GB

Blender Version
Broken:2.82 Beta 2020-01-22 12:36 Hash:a60a623a1ac5
Worked: (optional)

Short description of error
Adding very weak emissions to the scene weakens the highlights in cycles render

Exact steps for others to reproduce the error
Light_test.mp4
Uploaded blend file has strong emission, very weak emission and background light.
Very weak emissions have keyframes, 1 frame has a value of 0, and 2 frame have a value of 0.001.
Comparing the rendered images shows that the highlights of the 2 frame is weaker than the highlights of the 1 frame.
It seems strange that the highlights get weaker if you add a very weak emission that does not affect the scene.
Not included in the video, but placing very weak emissions close to strong emissions has the same effect.
If you add weak radiation, I think it will be a bit brighter, but the result will be darker.
Light_test.blend

**System Information** Operating system:macOS 10.15.2 Graphics card:Radeon Pro 580X 8 GB **Blender Version** Broken:2.82 Beta 2020-01-22 12:36 Hash:a60a623a1ac5 Worked: (optional) **Short description of error** Adding very weak emissions to the scene weakens the highlights in cycles render **Exact steps for others to reproduce the error** [Light_test.mp4](https://archive.blender.org/developer/F8295561/Light_test.mp4) Uploaded blend file has strong emission, very weak emission and background light. Very weak emissions have keyframes, 1 frame has a value of 0, and 2 frame have a value of 0.001. Comparing the rendered images shows that the highlights of the 2 frame is weaker than the highlights of the 1 frame. It seems strange that the highlights get weaker if you add a very weak emission that does not affect the scene. Not included in the video, but placing very weak emissions close to strong emissions has the same effect. If you add weak radiation, I think it will be a bit brighter, but the result will be darker. [Light_test.blend](https://archive.blender.org/developer/F8295559/Light_test.blend)
Author

Added subscriber: @cubimation

Added subscriber: @cubimation

Added subscriber: @MarcinTwarowski

Added subscriber: @MarcinTwarowski

It would be easier to test it if you simplified that scene. There are 88 thousand objects. I tried to delete the layer with those cubes but it takes Blender forever.

It would be easier to test it if you simplified that scene. There are 88 thousand objects. I tried to delete the layer with those cubes but it takes Blender forever.
Author

I upload a light sized file
Light_test2.blend

I upload a light sized file [Light_test2.blend](https://archive.blender.org/developer/F8297199/Light_test2.blend)

Added subscriber: @mano-wii

Added subscriber: @mano-wii

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

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

I can confirm.
Operating system: Windows-10-10.0.18941 64 Bits
Graphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.13559 Core Profile Context 26.20.12028.2

The problem is seen in both GPU and CPU

I can confirm. **Operating system:** Windows-10-10.0.18941 64 Bits **Graphics card:** Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.13559 Core Profile Context 26.20.12028.2 The problem is seen in both GPU and CPU
Member

Added subscriber: @EAW

Added subscriber: @EAW
Member

I don't think that this is a bug, but a known limitation of Monte Carlo ray tracing.
Cycles has 2 sampling methods.

Path Tracing
The Path Tracing integrator is a pure path tracer; at each hit it will bounce light in one direction and pick one light to receive lighting from.
Branched Path Tracing
The non-progressive Branched Path Tracing integrator offers finer control over sampling. It is similar to Path Tracing, but at the first hit it will split the path for different surface components and will take all lights into account for shading instead of just one.

You have a light with a strength of 678.4, the world background at 1, and then you add a light with a strength of 0.001. Frame 1: Ever pixel sample is either lit by the 678.4 light or the world background at 1, chosen at random. Frame 2: you add in another light source at 0.001. The result is that it will bring the overall lighting intensity down, because it is using up the random samples that would have gone to the other light sources.

If you switch to Branched Path Tracing, which takes all lights into account, you will see that the additional light makes the scene brighter.


Emissive materials have Multiple Importance Sampling on by default.

Materials with emission shaders can be configured to use Multiple Importance Sampling (Material Settings). This means that they will get rays sent directly towards them, rather than ending up there based on rays randomly bouncing around. For very bright mesh light sources, this can reduce noise significantly. However, when the emission is not particularly bright, this will take samples away from other brighter light sources for which it is important to find them this way.

Turn off Multiple Importance Sampling for the 0.001 light, even when using Path Tracing, and Frame 1 & 2 are identical, as the 0.001 light is too minuscule to contribute any change.

I don't think that this is a bug, but a known limitation of Monte Carlo ray tracing. Cycles has 2 [sampling methods](https://docs.blender.org/manual/en/2.81/render/cycles/render_settings/sampling.html). > **Path Tracing** >The Path Tracing integrator is a pure path tracer; at each hit it will bounce light in one direction and **pick one light to receive lighting from.** > **Branched Path Tracing** > The non-progressive Branched Path Tracing integrator offers finer control over sampling. It is similar to Path Tracing, but at the first hit it will split the path for different surface components and **will take all lights into account for shading instead of just one.** You have a light with a strength of 678.4, the world background at 1, and then you add a light with a strength of 0.001. Frame 1: Ever pixel sample is either lit by the 678.4 light or the world background at 1, chosen at random. Frame 2: you add in another light source at 0.001. The result is that it will bring the overall lighting intensity down, because it is using up the random samples that would have gone to the other light sources. If you switch to Branched Path Tracing, which takes all lights into account, you will see that the additional light makes the scene brighter. --- Emissive materials have [Multiple Importance Sampling](https://docs.blender.org/manual/en/2.81/render/cycles/optimizations/reducing_noise.html#multiple-importance-sampling) on by default. > Materials with emission shaders can be configured to use Multiple Importance Sampling (Material Settings). This means that they will get rays sent directly towards them, rather than ending up there based on rays randomly bouncing around. For very bright mesh light sources, this can reduce noise significantly. **However, when the emission is not particularly bright, this will take samples away from other brighter light sources for which it is important to find them this way.** Turn off Multiple Importance Sampling for the 0.001 light, even when using Path Tracing, and Frame 1 & 2 are identical, as the 0.001 light is too minuscule to contribute any change.

Added subscriber: @brecht

Added subscriber: @brecht
Member

Added subscriber: @Alaska

Added subscriber: @Alaska
Member

This appears to be the same issue as #97305 (As light object numbers increase, the indirect brightness of bright lights is decreased while indirect light clamping enabled - Cycles). I will merge these reports.

This appears to be the same issue as #97305 (As light object numbers increase, the indirect brightness of bright lights is decreased while indirect light clamping enabled - Cycles). I will merge these reports.
Member

Closed as duplicate of #97305

Closed as duplicate of #97305
Sign in to join this conversation.
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset Browser
Interest
Asset Browser Project Overview
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
Interest
EEVEE & Viewport
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
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
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
EEVEE & Viewport
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
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Priority
High
Priority
Low
Priority
Normal
Priority
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
5 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#73328
No description provided.