4.0 cycles spotlights produce unwanted refraction and glare #113693
Labels
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset System
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Code Documentation
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
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 & 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
Viewport & EEVEE
Interest
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
Legacy
Asset Browser Project
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
Asset System
Module
Core
Module
Development Management
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Module
Viewport & EEVEE
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Severity
High
Severity
Low
Severity
Normal
Severity
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
No due date set.
Dependencies
No dependencies set.
Reference: blender/blender#113693
Loading…
Reference in New Issue
Block a user
No description provided.
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?
System Information
Operating system: Windows-10-10.0.22621-SP0 64 Bits
Graphics card: Intel(R) Arc(TM) A770 Graphics Intel 4.6.0 - Build 31.0.101.4887
Blender Version
Broken: 4.0
Worked: 3.6 and earlier
the attached screenshot shows the same scene in 4 blender versions: 3.3, 3.5 (top left/top right), 3.6, 4.0 (bottom left/bottom right)
in 4.0, spotlights that were previously not visible are producing glare and refraction even outside of their cone of visibility. (ignore the top row of images.)
on the right bottom side of the case, a spotlight that overlaps with the case is illuminating the case brightly within its radius despite that area not being within the spotlight cone. this is direct illumination (there is no material on the case, and no glossy reflection).
on the right top of side of the case, the light reflects on the glass, also despite not being within the spotlight cone.
last, if you redirect the spotlight to point through the glass, the spotlight will be visible refracting/transmitting through the glass even when the camera is not within the cone. previous versions of blender would only show the light if the camera/viewpoint was directly in the cone.
these make it harder to avoid bright refractions/reflections in glass materials...this scene and another one i authored in 3.6 look markedly different (and worse) in 4.0 with bright reflections and refractions that i had avoided with deliberate light placement and cone size/angles. the refraction issue #3 is the worst one of the lot imho. the other 2 are easier to work around.
The release notes mention some changes to the lamps:
https://wiki.blender.org/wiki/Reference/Release_Notes/4.0/Rendering#Lights
I believe that the changes to Spot lamps are related to this commit:
#109329 Cycles: replace spot light disk sampling with sphere sampling
But in fact this compatibility breakage is strange. Didn't it deserve versioning?
@weizhen can know the details better.
Weizhen is currently away and will be for a few (5 or 6) more months. It's probably better to ask someone else like @brecht to comment on this until they're back.
This is a case of bad visualization. In Cycles the spotlight’s “cone of visibility” has always been wider than the overlay cone. For 4.0 you can imagine a light bulb with a lamp shade, before 4.0 the sampling was quite weird, it’s hard to visualize the shape, but also definitely larger than the cone you see.
The cone is drawn that way partly because overlay is handled by @fclem and EEVEE doesn’t support spotlight radius. We should be able to do better, either use different overlay depending on the render engine used, or change the overlay and support spotlight radius in EEVEE?
Besides, it should be expected that light with larger radius would illuminate a larger radius. If you keep the radius zero the illumination will stay in that cone, and you can change the blend if you just want it to look softer.
The change appears to be intentional and an improvement.
However, there is still a question: can this alteration in the rendering result be addressed through versioning? (Automatically setting a radius zero and maybe slightly increasing the angle of the cones for older files)
It can't be fixed in versioning, the radius also affects the softness of shadow. Which in most setups where the sphere does not overlap other objects is the more important thing to preserve.
There's some work going on to standardize these light shapes for USD, and one interesting behavior I saw in some other renderers is to have this work as a disk light that cuts off the end of the cone, which looks more natural.
In any case, I don't think this is a bug, even we if we can still improve the visualization and/or shape in the future.