blender
/
blender
134
398
Fork 580
Code Issues 6.3k Pull Requests 753 Projects 19 Wiki Activity

bugfix/usd_material_color_space #113374

Merged
Michael Kowalski merged 12 commits from CharlesWardlaw/blender:bugfix/usd_material_color_space into main 2023-12-06 18:18:07 +01:00
Conversation 12 Commits 12 Files Changed 2 +74 -28
Charles Wardlaw commented 2023-10-07 00:26:56 +02:00
Member
Copy Link

Many assets with UsdPreviewSurface materials exist in which the textures do not have authored color space settings. Because of this, assets with normal maps and other non-color data look incorrect on import because these textures are marked as being sRGB instead of Non-Color.

This patch fixes the color space choices for these textures based on connections to the PBR shader; if no color space is specified, non-Albedo maps get an automatic Non-Color assignment.

Testing: I used the USDZ file available here: https://www.metmuseum.org/-/media/ar-files/mma-ar_1979_206_380_wooddiety_audio_v1_final_max.usdz?la=en&hash=586A61E96C9C5237A4751584BAFC32A8

Before the patch:
image

After the patch:
image

Many assets with UsdPreviewSurface materials exist in which the textures do not have authored color space settings. Because of this, assets with normal maps and other non-color data look incorrect on import because these textures are marked as being sRGB instead of Non-Color. This patch fixes the color space choices for these textures based on connections to the PBR shader; if no color space is specified, non-Albedo maps get an automatic Non-Color assignment. Testing: I used the USDZ file available here: https://www.metmuseum.org/-/media/ar-files/mma-ar_1979_206_380_wooddiety_audio_v1_final_max.usdz?la=en&hash=586A61E96C9C5237A4751584BAFC32A8 Before the patch: ![image](/attachments/184f769a-0dbd-4ce2-a45d-81893f60a6d8) After the patch: ![image](/attachments/5fee42f9-3571-41ba-9ff0-f6e54d885abf)
image.png
497 KiB
image.png
788 KiB
Charles Wardlaw added 6 commits 2023-10-07 00:27:10 +02:00
Charles Wardlaw requested review from Michael Kowalski 2023-10-07 00:27:28 +02:00
Charles Wardlaw added this to the USD project 2023-10-07 00:27:34 +02:00
Charles Wardlaw requested review from Hans Goudey 2023-10-11 15:42:33 +02:00
Charles Wardlaw added 1 commit 2023-10-11 15:43:45 +02:00
Matt McLin commented 2023-10-23 20:11:05 +02:00
Member
Copy Link

I pulled this PR and built locally, can confirm this is working well for me. I also confirmed that my asset round-tripped OK.

I pulled this PR and built locally, can confirm this is working well for me. I also confirmed that my asset round-tripped OK.
Pablo Delgado commented 2023-10-24 15:44:53 +02:00
Contributor
Copy Link

Many assets [...] do not have authored color space settings.

Which is a feature, with the default value of UsdUVTexture:sourceColorSpace being specified as 'auto'.

This patch fixes the color space choices for these textures based on connections to the PBR shader [...]

I'm afraid this violates the UsdPreviewSurface specification which states:

auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture.

I think that the above behaviour is caused by two issues that do not get addressed by this PR.

  1. We do not implement 'auto', as Michael's comment states.
  2. The asset is just faulty. It should have 'sourceColorSpace:raw' authored on its normal maps, which is not the case.
> Many assets [...] do not have authored color space settings. Which is a feature, with the default value of UsdUVTexture:sourceColorSpace being specified as 'auto'. > This patch fixes the color space choices for these textures based on connections to the PBR shader [...] I'm afraid this violates the [UsdPreviewSurface specification](https://openusd.org/release/spec_usdpreviewsurface.html#texture-reader) which states: > auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture. I think that the above behaviour is caused by two issues that do not get addressed by this PR. 1. We do not implement 'auto', as [Michael's comment](https://projects.blender.org/blender/blender/src/commit/92a19747f024882f21b6c9eee2edc6ff35b3494d/source/blender/io/usd/intern/usd_reader_material.cc#L788-L791) states. 2. The asset is just faulty. It should have 'sourceColorSpace:raw' authored on its normal maps, which is not the case.
Michael Kowalski commented 2023-10-24 22:20:49 +02:00
Member
Copy Link

This patch fixes the color space choices for these textures based on connections to the PBR shader [...]

I'm afraid this violates the UsdPreviewSurface specification which states:

auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture.

I think that the above behaviour is caused by two issues that do not get addressed by this PR.

  1. We do not implement 'auto', as Michael's comment states.
  2. The asset is just faulty. It should have 'sourceColorSpace:raw' authored on its normal maps, which is not the case.

This occurred to me as well. It would be straightforward enough to check for sRGB color space with the following API:

#include <pxr/imaging/hio/image.h>

pxr::HioImageSharedPtr image_ptr = pxr::HioImage::OpenForReading(im_file);
bool isSRGB = image_ptr->IsColorSpaceSRGB();

However, the example asset would still be displayed incorrectly based on this test. Consequently, as you point out, the normal map should be authored with color space "raw".

To be honest, I'm a little on the fence about this, as it appears some DCCs do try to infer the color space, even if the asset does not strictly adhere to the standard.

@Matt-McLin @wave Do you have an opinion about this? Should we strictly adhere to the standard in this case and not try to infer the color space?

> > This patch fixes the color space choices for these textures based on connections to the PBR shader [...] > > I'm afraid this violates the [UsdPreviewSurface specification](https://openusd.org/release/spec_usdpreviewsurface.html#texture-reader) which states: > > auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture. > > I think that the above behaviour is caused by two issues that do not get addressed by this PR. > 1. We do not implement 'auto', as [Michael's comment](https://projects.blender.org/blender/blender/src/commit/92a19747f024882f21b6c9eee2edc6ff35b3494d/source/blender/io/usd/intern/usd_reader_material.cc#L788-L791) states. > 2. The asset is just faulty. It should have 'sourceColorSpace:raw' authored on its normal maps, which is not the case. This occurred to me as well. It would be straightforward enough to check for sRGB color space with the following API: ``` #include <pxr/imaging/hio/image.h> pxr::HioImageSharedPtr image_ptr = pxr::HioImage::OpenForReading(im_file); bool isSRGB = image_ptr->IsColorSpaceSRGB(); ``` However, the example asset would still be displayed incorrectly based on this test. Consequently, as you point out, the normal map should be authored with color space "raw". To be honest, I'm a little on the fence about this, as it appears some DCCs do try to infer the color space, even if the asset does not strictly adhere to the standard. @Matt-McLin @wave Do you have an opinion about this? Should we strictly adhere to the standard in this case and not try to infer the color space?
Michael Kowalski commented 2023-10-25 00:51:34 +02:00
Member
Copy Link

I wonder if "Emission Color" should also be considered a color input and default to sRGB.

I wonder if "Emission Color" should also be considered a color input and default to sRGB.
Pablo Delgado commented 2023-10-25 10:34:01 +02:00
Contributor
Copy Link

Yes, both emissiveColor and specularColor should be handled the same way as they are of type color3f.

Yes, both `emissiveColor` and `specularColor` should be handled the same way as they are of type color3f.
Michael Kowalski added the
Interest
USD
 label 2023-11-16 19:17:14 +01:00
Michael Kowalski commented 2023-11-20 23:06:32 +01:00
Member
Copy Link

I just wanted to resume the discussion. In my opinion, it's acceptable to infer the color space, as is being done in this pull request, even if that doesn't strictly adhere to the specification.

Are there objections to accepting this approach?

I just wanted to resume the discussion. In my opinion, it's acceptable to infer the color space, as is being done in this pull request, even if that doesn't strictly adhere to the specification. Are there objections to accepting this approach?
Michael Kowalski commented 2023-11-20 23:08:43 +01:00
Member
Copy Link

@deadpin @Matt-McLin Please let me know if you have an opinion about this. Thanks!

@deadpin @Matt-McLin Please let me know if you have an opinion about this. Thanks!
Jesse Yurkovich commented 2023-11-21 01:25:49 +01:00
Member
Copy Link

Given that usdview is also "broken" for this asset I'm not sure. Part of why USD exists is to remove the ambiguity of older formats and robustly interchange scenes. Going out of our way to perpetuate a broken ecosystem won't help that goal even if a, probably under staffed, museum got caught up in the mess.

If the spec isn't helping make a decision here I'd at least want an issue filed on USDPreviewSurface so we can reference it in a comment as part of any checkin we do.

It's a slippery slope to follow even if it's an easy fix on our side.

Similarly, the use of, yet another, 3rd form of the "raw" token should be removed from the PR at least. As a reviewer I'm now left wondering why we don't do this for every other token we have or if we need even more combinations etc. These fixups have a cost and, if we do add them, they should be done separately so they're easily reverted if necessary.

Given that `usdview` is also "broken" for this asset I'm not sure. Part of why USD exists is to remove the ambiguity of older formats and robustly interchange scenes. Going out of our way to perpetuate a broken ecosystem won't help that goal even if a, probably under staffed, museum got caught up in the mess. If the spec isn't helping make a decision here I'd at least want an issue filed on USDPreviewSurface so we can reference it in a comment as part of any checkin we do. It's a slippery slope to follow even if it's an easy fix on our side. Similarly, the use of, yet another, 3rd form of the "raw" token should be removed from the PR at least. As a reviewer I'm now left wondering why we don't do this for every other token we have or if we need even more combinations etc. These fixups have a cost and, if we do add them, they should be done separately so they're easily reverted if necessary.
👍 1
Matt McLin commented 2023-11-27 18:12:20 +01:00
Member
Copy Link

@makowalski Sorry for the late replies here. @wave, Dhruv, and I discussed this PR several weeks ago and all agreed that it should move forward, specifically regarding treating the auto/non-specified colorspaces for normal inputs as non-color. We did not discuss this 3rd form of "raw", I don't have an opinion on this. We also did not specifically discuss metallic/roughness inputs, but I believe it makes sense there as well.

Note that I disagree that this violates the UsdPreviewSurface specification. In spite of this phrasing:

auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture.

There is also this:

            Normal map data is commonly expected to be linearly encoded.
            However, many image-writing tools automatically set the profile
            of three-channel, 8-bit images to SRGB. To prevent an unwanted
            transformation, the sourceColorSpace must also be set to "raw".

The above is acknowledging that it is never correct for the normal inputs to be anything other than "raw", regardless of whether the source is 8-bit or seems to be SRGB. Thus it is always a mistake on the part of the asset author or authoring tool if this is anything other than "raw", and I believe this means we are doing a great service to the community by working around these mistakes on import, of which we have seen many examples in the wild.

@makowalski Sorry for the late replies here. @wave, Dhruv, and I discussed this PR several weeks ago and all agreed that it should move forward, specifically regarding treating the auto/non-specified colorspaces for normal inputs as non-color. We did not discuss this 3rd form of "raw", I don't have an opinion on this. We also did not specifically discuss metallic/roughness inputs, but I believe it makes sense there as well. Note that I disagree that this violates the UsdPreviewSurface specification. In spite of this phrasing: ``` auto : Check for gamma/color space metadata in the texture file itself; if metadata is indicative of sRGB, mark texture as sRGB . If no relevant metadata is found, mark texture as sRGB if it is either 8-bit and has 3 channels or if it is 8-bit and has 4 channels. Otherwise, do not mark texture as sRGB and use texture data as it was read from the texture. ``` There is also this: ``` Normal map data is commonly expected to be linearly encoded. However, many image-writing tools automatically set the profile of three-channel, 8-bit images to SRGB. To prevent an unwanted transformation, the sourceColorSpace must also be set to "raw". ``` The above is acknowledging that it is _never_ correct for the normal inputs to be anything other than "raw", regardless of whether the source is 8-bit or seems to be SRGB. Thus it is always a mistake on the part of the asset author or authoring tool if this is anything other than "raw", and I believe this means we are doing a great service to the community by working around these mistakes on import, of which we have seen many examples in the wild.
👍 1
Matt McLin commented 2023-11-27 18:17:33 +01:00
Member
Copy Link

@deadpin While comparing to usdview is a very reasonable thing in general, I have imagined the aim for Blender's importer functionality is different than for usdview. Probably we don't want Blender to be a reference tool to check the correctness of assets authored elsewhere? I think a focus on "it just works" is likely more beneficial to the broader Blender community here.

@deadpin While comparing to usdview is a very reasonable thing in general, I have imagined the aim for Blender's importer functionality is different than for usdview. Probably we don't want Blender to be a reference tool to check the correctness of assets authored elsewhere? I think a focus on "it just works" is likely more beneficial to the broader Blender community here.
Charles Wardlaw added 1 commit 2023-12-01 00:00:07 +01:00
Jesse Yurkovich requested changes 2023-12-01 06:44:41 +01:00
Jesse Yurkovich left a comment
Member
Copy Link

Alright, based on the meeting this morning, let's do this :) Can you merge in main as I think there's conflicts with your other recent change that was just checked in.

Alright, based on the meeting this morning, let's do this :) Can you merge in main as I think there's conflicts with your other recent change that was just checked in.
source/blender/io/usd/intern/usd_reader_material.cc Outdated
@ -801,0 +820,4 @@
color_space = usdtokens::auto_;
}
if (ELEM(color_space, usdtokens::auto_)) {
Jesse Yurkovich commented 2023-12-01 06:41:58 +01:00
Member
Copy Link

When comparing to just 1 item, you don't need to use the ELEM macro. Here and 1 other place just below this.

When comparing to just 1 item, you don't need to use the ELEM macro. Here and 1 other place just below this.
CharlesWardlaw marked this conversation as resolved
source/blender/io/usd/intern/usd_reader_material.h Outdated
@ -102,2 +102,3 @@
int column,
NodePlacementContext *r_ctx) const;
NodePlacementContext *r_ctx,
bool is_color_corrected = false) const;
Jesse Yurkovich commented 2023-12-01 06:43:59 +01:00
Member
Copy Link

The blender style guide says to keep the r_ parameters grouped last but that will obviously conflict with the optional param. I think I'm fine with just making this a required parameter all-up even if that causes the set_node_input calls to be a bit more verbose. set_node_input is really the only func that makes use of the optional-ness anyhow as the other functions will always provide the param in some form.

The blender style guide says to keep the `r_` parameters grouped last but that will obviously conflict with the optional param. I think I'm fine with just making this a required parameter all-up even if that causes the `set_node_input` calls to be a bit more verbose. `set_node_input` is really the only func that makes use of the optional-ness anyhow as the other functions will always provide the param in some form.
CharlesWardlaw marked this conversation as resolved
Charles Wardlaw added 1 commit 2023-12-05 19:21:35 +01:00
Charles Wardlaw added 1 commit 2023-12-05 19:26:01 +01:00
Charles Wardlaw added 1 commit 2023-12-05 19:48:49 +01:00
buildbot/vexp-code-patch-coordinator Build done. Details
72fd523518 Merge fixes
Jesse Yurkovich approved these changes 2023-12-05 20:39:54 +01:00
Michael Kowalski approved these changes 2023-12-06 15:37:47 +01:00
Michael Kowalski commented 2023-12-06 15:40:38 +01:00
Member
Copy Link

@blender-bot build

@blender-bot build
Michael Kowalski commented 2023-12-06 16:25:20 +01:00
Member
Copy Link

@CharlesWardlaw Would you mind running make format on the code, please?

Otherwise, the changes look good to me. Thank you!

@CharlesWardlaw Would you mind running `make format` on the code, please? Otherwise, the changes look good to me. Thank you!
Charles Wardlaw added 1 commit 2023-12-06 17:50:57 +01:00
Michael Kowalski merged commit b7fa45b68d into main 2023-12-06 18:18:07 +01:00
Sign in to join this conversation.
Reviewers
No reviewers
Hans Goudey
Jesse Yurkovich
Michael Kowalski
Labels
Clear labels   
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
This issue affects/is about backward or forward 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
Issues relating to Wayland windowing support.

  
Interest
Workbench

  
Interest: X11

Issues relating to Xorg/X11 support.

  
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
Issues relating to security: https://wiki.blender.org/wiki/Process/Vulnerability_Reports

  
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 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
Milestone
Clear milestone
No items
No Milestone
Projects
Clear projects
No project
USD
Assignees
Clear assignees
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#113374
No description provided.
Delete Branch "CharlesWardlaw/blender:bugfix/usd_material_color_space"

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?