Break dependency from `bevy_anti_alias` on `bevy_post_process` by splitting `PostProcess` set by Zeophlite · Pull Request #23098 · bevyengine/bevy

Zeophlite · 2026-02-21T15:37:29Z

Objective

During Replace RenderGraph with systems #22144 , we slightly linearlised the crate build, by explicit dependency on system ordering. See this comment

Solution

Split PostProcess set into EarlyPostProcess and PostProcess sets
Move taa and dlss into EarlyPostProcess

Testing

CI
cargo run --example anti_aliasing

github-actions · 2026-02-21T15:45:33Z

Your PR caused a change in the graphical output of an example or rendering test. This might be intentional, but it could also mean that something broke!
You can review it at https://pixel-eagle.com/project/B04F67C0-C054-4A6F-92EC-F599FEC2FD1D?filter=PR-23098

If it's expected, please add the M-Deliberate-Rendering-Change label.

If this change seems unrelated to your PR, you can consider updating your PR to target the latest main branch, either by rebasing or merging main into it.

github-actions · 2026-02-21T16:27:06Z

Your PR caused a change in the graphical output of an example or rendering test. This might be intentional, but it could also mean that something broke!
You can review it at https://pixel-eagle.com/project/B04F67C0-C054-4A6F-92EC-F599FEC2FD1D?filter=PR-23098

If it's expected, please add the M-Deliberate-Rendering-Change label.

If this change seems unrelated to your PR, you can consider updating your PR to target the latest main branch, either by rebasing or merging main into it.

atlv24

I think the naming of the new sets needs some more thinking. Taa running in PostProcess doesnt make sense. How about
EarlyPostProcess
AntiAlias
PostProcess

So the anti aliasing actually runs in the AntiAlias set?

atlv24 · 2026-02-22T05:02:42Z

the anti aliasing example crashes on a dependency cycle:

thread '<unnamed>' (8214412) panicked at crates/bevy_ecs/src/schedule/schedule.rs:536:13:
Error when initializing schedule Core3d: schedule has 0 before/after cycle(s):

note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Encountered a panic in system `bevy_core_pipeline::schedule::camera_driver`!
Encountered a panic in system `bevy_render::renderer::render_system`!
Encountered a panic in system `bevy_render::run_render_schedule`!

Zeophlite · 2026-02-22T14:30:13Z

Definitely made a mistake with the ordering, anti_aliasing example worked

atlv24

We should probably put the rest of the anti aliasing solutions into the AntiAlias set

github-actions · 2026-02-24T04:06:16Z

Your PR caused a change in the graphical output of an example or rendering test. This might be intentional, but it could also mean that something broke!
You can review it at https://pixel-eagle.com/project/B04F67C0-C054-4A6F-92EC-F599FEC2FD1D?filter=PR-23098

If it's expected, please add the M-Deliberate-Rendering-Change label.

If this change seems unrelated to your PR, you can consider updating your PR to target the latest main branch, either by rebasing or merging main into it.

Zeophlite · 2026-02-24T04:34:25Z

Hmm, TAA flickers with anti_aliasing example now :(

In crates/bevy_anti_alias/src/taa/taa.wgsl there's

// TAA is ideally applied after tonemapping, but before post processing
// Post processing wants to go before tonemapping, which conflicts
// Solution: Put TAA before tonemapping, tonemap TAA input, apply TAA, invert-tonemap TAA output

I think this might be a little more complicated to complete

Zeophlite · 2026-03-03T11:11:20Z

Hmm, TAA flickers with anti_aliasing example now :(

In crates/bevy_anti_alias/src/taa/taa.wgsl there's
// TAA is ideally applied after tonemapping, but before post processing
// Post processing wants to go before tonemapping, which conflicts
// Solution: Put TAA before tonemapping, tonemap TAA input, apply TAA, invert-tonemap TAA output
I think this might be a little more complicated to complete

This is resolved my merging main

alice-i-cecile

I like the explicit AntiAlias step. Can you update the PR description and title?

kfc35

Just have some questions about the ordering of the systems that seem to have changed, but also I’m not too familiar with rendering so if I’m asking about dumb things (or worse, I’m incapable of reading system ordering 🙈 ), feel free to disregard

Zeophlite · 2026-03-17T03:38:18Z

@kfc35 Thanks for your review, this has made me check this PR from first principles.

My read of the relevant systems is as follows:

blue are bevy_anti_alias
red are bevy_post_process
grey are bevy_core_pipelines (which is the common ancestor)

The goal is to remove blue/red edges (all others are fine).

Therefore, we only really need two sets, EarlyPostProcess which has taa and dlss , and PostProcess which is the rest.

Graph generation

digraph G {

  // bevy_anti_alias = blue
  cas [color=blue];
  dlss [color=blue];
  fxaa [color=blue];
  smaa [color=blue];
  taa [color=blue];

  // bevy_post_process = red
  bloom [color=red];
  motion_blur [color=red];

  // other = grey
  tonemapping [color=grey];


  fxaa -> cas;         // crates/bevy_anti_alias/src/contrast_adaptive_sharpening/mod.rs
  dlss -> bloom;       // crates/bevy_anti_alias/src/dlss/mod.rs
  tonemapping -> fxaa  // crates/bevy_anti_alias/src/fxaa/mod.rs
  tonemapping -> smaa; // crates/bevy_anti_alias/src/smaa/mod.rs
  taa -> motion_blur;  // crates/bevy_anti_alias/src/taa/mod.rs
  taa -> bloom;        // crates/bevy_anti_alias/src/taa/mod.rs

  bloom -> tonemapping; // crates/bevy_post_process/src/bloom/mod.rs
  motion_blur -> bloom; // crates/bevy_post_process/src/motion_blur/mod.rs
}

…ng PostProcess set

kfc35

Nice, this looks great! Thanks for making the graph; I should keep https://graph.flyte.org/ handy too. This just motivates me to start thinking about automatic schedule visualizations for bevy systems...

tychedelia

Nice investigation, thanks.

# Objective Fixes #23472 (and a similar issue for wireframes) When `EarlyPostProcess` was created in #23098, some existing systems that were scheduled between `MainPass` and `PostProcess` weren't updated. This caused them to race with TAA, which causes flickering and other artifacts. ## Solution Order them relative to `EarlyPostProcess` instead. ## Testing `cargo run --example scrolling_fog` for volumetric fog. For wireframes you can see flickering on main if you add TAA to `wireframe.rs`: ```diff index 74376ab..4ac71fcdc 100644 --- a/examples/3d/wireframe.rs +++ b/examples/3d/wireframe.rs @@ -114,6 +114,8 @@ fn setup( commands.spawn(( Camera3d::default(), Transform::from_xyz(-2.0, 2.5, 5.0).looking_at(Vec3::ZERO, Vec3::Y), + Msaa::Off, + TemporalAntiAliasing::default(), )); ```

…itting `PostProcess` set (bevyengine#23098) # Objective - During bevyengine#22144 , we slightly linearlised the crate build, by explicit dependency on system ordering. [See this comment](https://github.com/bevyengine/bevy/pull/22144/changes#r2697878715) ## Solution - Split `PostProcess` set into `EarlyPostProcess` and `PostProcess` sets - Move `taa` and `dlss` into `EarlyPostProcess` <img width="309" height="539" alt="aa-no-pp" src="https://github.com/user-attachments/assets/bb32b71a-4743-46e4-ad2c-f06ef7db854f" /> ## Testing - CI - `cargo run --example anti_aliasing`

# Objective Fixes bevyengine#23472 (and a similar issue for wireframes) When `EarlyPostProcess` was created in bevyengine#23098, some existing systems that were scheduled between `MainPass` and `PostProcess` weren't updated. This caused them to race with TAA, which causes flickering and other artifacts. ## Solution Order them relative to `EarlyPostProcess` instead. ## Testing `cargo run --example scrolling_fog` for volumetric fog. For wireframes you can see flickering on main if you add TAA to `wireframe.rs`: ```diff index 74376ab..4ac71fcdc 100644 --- a/examples/3d/wireframe.rs +++ b/examples/3d/wireframe.rs @@ -114,6 +114,8 @@ fn setup( commands.spawn(( Camera3d::default(), Transform::from_xyz(-2.0, 2.5, 5.0).looking_at(Vec3::ZERO, Vec3::Y), + Msaa::Off, + TemporalAntiAliasing::default(), )); ```

github-actions Bot mentioned this pull request Feb 21, 2026

23098 bevyengine/bevy-example-runner#214

Closed

Zeophlite requested a review from atlv24 February 21, 2026 15:38

Zeophlite added A-Rendering Drawing game state to the screen D-Straightforward Simple bug fixes and API improvements, docs, test and examples S-Needs-Review Needs reviewer attention (from anyone!) to move forward labels Feb 21, 2026

github-project-automation Bot added this to Rendering Feb 21, 2026

github-project-automation Bot moved this to Needs SME Triage in Rendering Feb 21, 2026

atlv24 reviewed Feb 22, 2026

View reviewed changes

Zeophlite force-pushed the aa-no-pp branch 2 times, most recently from f47c8c1 to 78937aa Compare February 22, 2026 14:28

github-actions Bot mentioned this pull request Feb 22, 2026

23098 bevyengine/bevy-example-runner#216

Closed

atlv24 reviewed Feb 23, 2026

View reviewed changes

Comment thread crates/bevy_core_pipeline/src/schedule.rs Outdated

alice-i-cecile added S-Waiting-on-Author The author needs to make changes or address concerns before this can be merged and removed S-Needs-Review Needs reviewer attention (from anyone!) to move forward labels Feb 23, 2026

alice-i-cecile changed the title ~~bevy_anti_alias doesn't depend on bevy_post_process~~ Break dependency from bevy_anti_alias on bevy_post_process Feb 23, 2026

Zeophlite added S-Needs-Help The author needs help finishing this PR. and removed S-Waiting-on-Author The author needs to make changes or address concerns before this can be merged labels Feb 24, 2026

alice-i-cecile requested a review from tychedelia February 24, 2026 17:13

Zeophlite added this to the 0.19 milestone Feb 25, 2026

Zeophlite added S-Needs-Review Needs reviewer attention (from anyone!) to move forward and removed S-Needs-Help The author needs help finishing this PR. labels Mar 3, 2026

alice-i-cecile approved these changes Mar 3, 2026

View reviewed changes

Zeophlite changed the title ~~Break dependency from bevy_anti_alias on bevy_post_process~~ Break dependency from bevy_anti_alias on bevy_post_process by splitting PostProcess ser Mar 3, 2026

Zeophlite changed the title ~~Break dependency from bevy_anti_alias on bevy_post_process by splitting PostProcess ser~~ Break dependency from bevy_anti_alias on bevy_post_process by splitting PostProcess set Mar 3, 2026

kfc35 reviewed Mar 14, 2026

View reviewed changes

Comment thread crates/bevy_anti_alias/src/dlss/mod.rs

Comment thread crates/bevy_anti_alias/src/taa/mod.rs

Comment thread crates/bevy_post_process/src/motion_blur/mod.rs

alice-i-cecile added S-Waiting-on-Author The author needs to make changes or address concerns before this can be merged and removed S-Needs-Review Needs reviewer attention (from anyone!) to move forward labels Mar 14, 2026

Zeophlite marked this pull request as draft March 17, 2026 03:41

Zeophlite force-pushed the aa-no-pp branch 2 times, most recently from 354e44f to 927e0d5 Compare March 17, 2026 03:53

Zeophlite added S-Needs-Review Needs reviewer attention (from anyone!) to move forward and removed S-Waiting-on-Author The author needs to make changes or address concerns before this can be merged labels Mar 17, 2026

Zeophlite marked this pull request as ready for review March 17, 2026 03:54

Zeophlite force-pushed the aa-no-pp branch from 927e0d5 to fd75d90 Compare March 17, 2026 03:55

Break dependency from bevy_anti_alias on bevy_post_process by splitti…

4aea3d8

…ng PostProcess set

Zeophlite force-pushed the aa-no-pp branch from fd75d90 to 4aea3d8 Compare March 17, 2026 03:57

kfc35 approved these changes Mar 17, 2026

View reviewed changes

kfc35 added S-Ready-For-Final-Review This PR has been approved by the community. It's ready for a maintainer to consider merging it and removed S-Needs-Review Needs reviewer attention (from anyone!) to move forward labels Mar 17, 2026

tychedelia approved these changes Mar 17, 2026

View reviewed changes

alice-i-cecile added this pull request to the merge queue Mar 18, 2026

Merged via the queue into bevyengine:main with commit c36ea11 Mar 18, 2026
44 checks passed

github-project-automation Bot moved this from Needs SME Triage to Done in Rendering Mar 18, 2026

mockersf mentioned this pull request Mar 22, 2026

scrolling_fog example broken since PostProcess set split #23472

Closed

kristoff3r mentioned this pull request Mar 28, 2026

Fix PostProcess orderings for fog and wireframes #23558

Merged

Uh oh!

Conversation

Zeophlite commented Feb 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Objective

Solution

Testing

Uh oh!

github-actions Bot commented Feb 21, 2026

Uh oh!

github-actions Bot commented Feb 21, 2026

Uh oh!

atlv24 left a comment

Choose a reason for hiding this comment

Uh oh!

atlv24 commented Feb 22, 2026

Uh oh!

Zeophlite commented Feb 22, 2026

Uh oh!

atlv24 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

github-actions Bot commented Feb 24, 2026

Uh oh!

Zeophlite commented Feb 24, 2026

Uh oh!

Zeophlite commented Mar 3, 2026

Uh oh!

alice-i-cecile left a comment

Choose a reason for hiding this comment

Uh oh!

kfc35 left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Zeophlite commented Mar 17, 2026

Uh oh!

kfc35 left a comment

Choose a reason for hiding this comment

Uh oh!

tychedelia left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Zeophlite commented Feb 21, 2026 •

edited

Loading

kfc35 left a comment •

edited

Loading