Wrong colors in interface after HDR commit on macOS #111260

Closed
opened 2023-08-18 12:57:17 +02:00 by Sergey Sharybin · 2 comments

System Information
Operating system: macOS-13.4.1-arm64-arm-64bit 64 Bits
Graphics card: Metal API Apple M2 Max 1.2

Blender Version
Broken: version: 4.0.0 Alpha, branch: main, commit date: 2023-08-17 23:26, hash: 0fb4c107db92
Worked: Blender versions before the HDR commit 2367ed2ef2

Short description of error

The HDR commit seems to have introduced unexpected changes in the way colors are presented on display even when HDR option is disabled.

It is caused by the changes in the GHOST in the pull request #105662. Reverting those changes makes colors to be displayed the same as they used to be.

it is very easy to see when toggling between two Blender builds. I did not find a reliable way of making a screenshot. It seems finder makes them look the same, so I've re-saved to the screenshots in Blender. It allowed to demonstrate the difference as a screenshot.

Again, when toggling between two builds this trickery is not needed.

Before After
Before.png After.png

Exact steps for others to reproduce the error

  • Open attached file
  • Render in Blender builds before and after the HDR commit
  • Observe that after the HDR commit the red looks more orange

It is also possible to compare colors between Blender 3.6 and current 4.0 builsbot builds, but there seems to be some other changes done since 3.6 which does make colors ever so slightly different.

**System Information** Operating system: macOS-13.4.1-arm64-arm-64bit 64 Bits Graphics card: Metal API Apple M2 Max 1.2 **Blender Version** Broken: version: 4.0.0 Alpha, branch: main, commit date: 2023-08-17 23:26, hash: `0fb4c107db92` Worked: Blender versions before the HDR commit 2367ed2ef24 **Short description of error** The HDR commit seems to have introduced unexpected changes in the way colors are presented on display even when HDR option is disabled. It is caused by the changes in the GHOST in the pull request #105662. Reverting those changes makes colors to be displayed the same as they used to be. it is very easy to see when toggling between two Blender builds. I did not find a reliable way of making a screenshot. It seems finder makes them look the same, so I've re-saved to the screenshots in Blender. It allowed to demonstrate the difference as a screenshot. Again, when toggling between two builds this trickery is not needed. | Before | After | | -- | -- | | ![Before.png](/attachments/8ad04f43-1f04-4f4d-965d-4d2eeb9c446b) | ![After.png](/attachments/6770833c-c452-420d-8027-b730c0d97d47) | **Exact steps for others to reproduce the error** * Open attached file * Render in Blender builds before and after the HDR commit * Observe that after the HDR commit the red looks more orange It is also possible to compare colors between Blender 3.6 and current 4.0 builsbot builds, but there seems to be some other changes done since 3.6 which does make colors ever so slightly different.
Sergey Sharybin added the
Type
Report
Priority
Normal
Status
Needs Triage
labels 2023-08-18 12:57:18 +02:00
Iliya Katushenock added the
Interest
Compositing
label 2023-08-18 12:57:56 +02:00

I can also confirm on an Apple M1 Apple 4.1 Metal - 83.1

Poking @Jason-Fielder as he is the author of the commit (although the bug may be elsewhere)

I can also confirm on an Apple M1 Apple 4.1 Metal - 83.1 Poking @Jason-Fielder as he is the author of the commit (although the bug may be elsewhere)
Germano Cavalcante added the
Priority
High
label 2023-08-18 16:33:41 +02:00
Author
Owner

I was giving it a deeper investigation and comparison.
The long story short: the current state of the main branch actually appears to be correct!

Longer story below.

From the inspiration of the proposed code change and detective work Michael did comparing different Blender versions and Metal layer configuration the main observation is: prior to the HDR commit colors provided by Blender were considered to be in Display P3, which made the colors look nice saturated. And this kind of what skewed our visual perception.

Blender does provide the output colors in the sRGB space, and it is more narrow than the Display P3. So it is expected that some desaturation happens. In fact, this desaturation can be seen when saving pure render result in Blender 3.6 as PNG, and viewing it in an external viewer (like, finder viewer).

Surely, one might argue that particular viewer application might not do proper color space conversion. This is why I've also cross-referenced the on-screen result of the PNG and the latest Blender 4.0 main branch with several image/photo editing software, and all of them are in an agreement.

Additionally, Nathan helped me doing one extra test. Since he is on Linux and regular sRGB monitor we put MacBook next to his monitor and showed the pure red render result on both screens. Surely, we don't really have developers' monitors calibrated, but the Blender 3.6 was way too saturated, and the latest main branch looked much closer to what we've seen on the sRGB screen.

So after all this investigation it seems that it is not a bug, but rather expected change caused by proper color configuration.

I was giving it a deeper investigation and comparison. The long story short: the current state of the main branch actually appears to be correct! Longer story below. From the inspiration of the proposed code change and detective work Michael did comparing different Blender versions and Metal layer configuration the main observation is: prior to the HDR commit colors provided by Blender were considered to be in Display P3, which made the colors look nice saturated. And this kind of what skewed our visual perception. Blender does provide the output colors in the sRGB space, and it is more narrow than the Display P3. So it is expected that some desaturation happens. In fact, this desaturation can be seen when saving pure render result in Blender 3.6 as PNG, and viewing it in an external viewer (like, finder viewer). Surely, one might argue that particular viewer application might not do proper color space conversion. This is why I've also cross-referenced the on-screen result of the PNG and the latest Blender 4.0 main branch with several image/photo editing software, and all of them are in an agreement. Additionally, Nathan helped me doing one extra test. Since he is on Linux and regular sRGB monitor we put MacBook next to his monitor and showed the pure red render result on both screens. Surely, we don't really have developers' monitors calibrated, but the Blender 3.6 was way too saturated, and the latest main branch looked much closer to what we've seen on the sRGB screen. So after all this investigation it seems that it is not a bug, but rather expected change caused by proper color configuration.
Blender Bot added
Status
Archived
and removed
Status
Confirmed
labels 2023-08-23 14:56:06 +02:00
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
2 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#111260
No description provided.