Procedural atmospheric scattering #16314

ecoskey · 2024-11-09T19:55:59Z

Implement procedural atmospheric scattering from Sebastien Hillaire's 2020 paper. This approach should scale well even down to mobile hardware, and is physically accurate.

TODO

implement multiscattering (followup?)
handle scene units better
fix over-exposure
check physical correctness
check that coordinates are correct (lat-long -> ray direction specifically, and that signs for depth + view dir are correct)
integrate with pcwalton's volumetrics code (could have volumetric shadows with accurate sky color!!) (definitely in a followup)
draw sun disks (get radius from shadow softness value?)
handle hdr/non-hdr view texture
set camera depth texture usages automatically

Showcase

Check the example in examples/3d/atmosphere.rs. Currently is super over-exposed when applied to camera, but in a graphics debugger the LUTs themselves look fine.

- add lights to bind groups - add lut sample count to settings struct

and more things as well

mostly

- add lights to bind groups - add lut sample count to settings struct

and more things as well

mostly

github-actions · 2024-11-09T21:17:00Z

The generated examples/README.md is out of sync with the example metadata in Cargo.toml or the example readme template. Please run cargo run -p build-templated-pages -- update examples to update it, and commit the file change.

pcwalton · 2024-11-09T22:04:53Z

Wow, this is very impressive! I'll review.

aevyrie · 2024-11-09T22:49:10Z

Considering the Atmosphere takes parameters about the inner and outer radius of the atmosphere, maybe it would make sense if that component was on another entity? That would allow you to compute the camera's view within (or outside) the atmosphere based on the position of the atmosphere and the camera using their GlobalTransforms. Excited to see this implemented!

github-actions · 2024-11-10T02:02:57Z

The generated examples/README.md is out of sync with the example metadata in Cargo.toml or the example readme template. Please run cargo run -p build-templated-pages -- update examples to update it, and commit the file change.

github-actions · 2024-11-10T06:38:19Z

The generated examples/README.md is out of sync with the example metadata in Cargo.toml or the example readme template. Please run cargo run -p build-templated-pages -- update examples to update it, and commit the file change.

pcwalton · 2024-11-10T18:03:29Z

Would you like reviews/feedback on this at this stage or is it too early?

pcwalton · 2024-11-10T18:07:07Z

crates/bevy_pbr/src/atmosphere/functions.wgsl

+fn sample_sky_view_lut(ray_dir: vec3<f32>) -> vec3<f32> {
+    let lat_long = ray_dir_to_lat_long(ray_dir);
+    let uv = sky_view_lut_lat_long_to_uv(lat_long.x, lat_long.y);
+    return textureSampleLevel(sky_view_lut, sky_view_lut_sampler, uv, 0.0).rgb;


Inverse trigonometric functions are really slow on GPU (and on CPU). Can you use a cubemap LUT instead? Directional data is basically why cubemap textures exist.

I believe the author did it this way because the latitude is parameterized to put most of the detail near the horizon, which does have a noticeable effect on the final result. There's probably room to speed up my math here, but I'd want to make sure a cubemap could preserve that detail before switching.

I should check how Hillaire implements the conversion, since apparently it's good enough for unreal

pcwalton · 2024-11-10T18:08:56Z

crates/bevy_pbr/src/atmosphere/functions.wgsl

+}
+
+// Convert uv [0.0 .. 1.0] coordinate to ndc space xy [-1.0 .. 1.0]
+fn uv_to_ndc(uv: vec2<f32>) -> vec2<f32> {


This function already exists under that name in view_transformations.wgsl

I copied it because I wanted to avoid the import bringing in all the mesh view bindings, which would overlap with atmosphere/bindings.wgsl. Maybe I'm misunderstanding how the imports work though

pcwalton · 2024-11-10T18:09:23Z

crates/bevy_pbr/src/atmosphere/functions.wgsl

+}
+
+/// Convert a ndc space position to world space
+fn position_ndc_to_world(ndc_pos: vec3<f32>) -> vec3<f32> {


This function already exists under that name in view_transformations.wgsl

I copied it because I wanted to avoid the import bringing in all the mesh view bindings, which would overlap with atmosphere/bindings.wgsl. Maybe I'm misunderstanding how the imports work though

pcwalton · 2024-11-10T18:10:34Z

crates/bevy_pbr/src/atmosphere/functions.wgsl

+    return FRAC_3_16_PI * (1 + (neg_LdotV * neg_LdotV));
+}
+
+fn henyey_greenstein(neg_LdotV: f32) -> f32 {


Probably worth factoring this out somewhere so that both volumetric fog and atmosphere can use it

pcwalton · 2024-11-10T18:12:10Z

crates/bevy_pbr/src/atmosphere/sky_view_lut.wgsl

+    let horizontal_rotation = mat2x2(cos_long, -sin_long, sin_long, cos_long);
+    let horizontal = horizontal_rotation * vec2(-view.world_from_view[2].xz);
+
+    return normalize(vec3(horizontal.x, sin(lat_long.x), horizontal.y));


Yeah, this would definitely be simpler as a cubemap.

ecoskey · 2024-11-10T18:15:20Z

Would you like reviews/feedback on this at this stage or is it too early?

The structure of everything is pretty much final so feedback on that would be nice :). At this stage I'm mainly trying to make everything physically correct, so things will change there but if you see something I'm clearly doing wrong lmk

ecoskey · 2024-11-10T20:39:43Z

Considering the Atmosphere takes parameters about the inner and outer radius of the atmosphere, maybe it would make sense if that component was an another entity? That would allow you to compute the camera's view within (or outside) the atmosphere based on the position of the atmosphere and the camera using their GlobalTransforms. Excited to see this implemented!

Some of the main assumptions made are that the center of the planet is at position (view.x, -bottom_radius, view.z) and that the camera never leaves the atmosphere. So this implementation doesn't support worlds that are actually spherical, or space scenes where the camera can view the atmosphere from above. In the future once this gets integrated with pcwalton's raytracing code, it might be worth removing these restrictions. In that case I'd agree that the atmosphere/planet should be its own entity

github-actions · 2024-11-10T20:45:22Z

The generated examples/README.md is out of sync with the example metadata in Cargo.toml or the example readme template. Please run cargo run -p build-templated-pages -- update examples to update it, and commit the file change.

github-actions · 2024-11-11T03:52:27Z

You added a new feature but didn't update the readme. Please run cargo run -p build-templated-pages -- update features to update it, and commit the file change.

aevyrie · 2024-11-11T18:11:33Z

Some of the main assumptions made

I thought the Hillaire paper explicitly allows for space views of the atmosphere? Anyway, I totally understand if it is out of scope. Thanks for working on this!

ecoskey · 2024-11-12T18:09:36Z

Some of the main assumptions made

I thought the Hillaire paper explicitly allows for space views of the atmosphere? Anyway, I totally understand if it is out of scope. Thanks for working on this!

It does, but the higher altitude the camera is the worse precision the sky-view and aerial-view LUTs are, even within the atmosphere. The paper recommends switching to per-pixel raytracing for high-altitude or space shots, which is where I'd want to integrate pcwalton's volumetrics code. I'd prefer that in a followup though :)

ecoskey added 30 commits August 29, 2024 09:58

wip

9d948ab

WIP

515dfde

Merge branch 'main' into proc_sky

ceb4af1

WIP

534f505

use buffer binding

3fa0980

WIP + example

d5c58f0

bind groups and more prep

28bd632

WIP

4c3de51

- add lights to bind groups - add lut sample count to settings struct

add ground albedo to bind groups

e18baec

add apply pass

823a338

and more things as well

bwah

222b11d

WIP + finished aerial view lut

39c1075

small fixes

a682732

update bindings

fdc86c5

mostly

Merge branch 'main' into proc_sky

6f3a7c5

WIP

0096c59

use buffer binding

bad3819

WIP + example

c1c73a1

bind groups and more prep

6ed73d4

WIP

1b215d8

- add lights to bind groups - add lut sample count to settings struct

add ground albedo to bind groups

6050cb3

add apply pass

3c4420f

and more things as well

bwah

98285f3

WIP + finished aerial view lut

5cabba8

small fixes

6e9251b

update bindings

33ded2f

mostly

Merge branch 'proc_sky' of github.com:ecoskey/bevy into proc_sky

40a9b54

sky_view_lut progress

7f05f5d

fix bindings and such

9be3e7b

misc fixes

7f1ed99

alice-i-cecile added M-Needs-Release-Note Work that should be called out in the blog due to impact S-Waiting-on-Author The author needs to make changes or address concerns before this can be merged labels Nov 9, 2024

Merge branch 'main' into proc_sky

ac1ee5b

more small progress

d03c2bb

format toml

490cc5b

BenjaminBrienen added the D-Complex Quite challenging from either a design or technical perspective. Ask for help! label Nov 10, 2024

pcwalton reviewed Nov 10, 2024

View reviewed changes

Merge branch 'main' into proc_sky

ac4663d

ecoskey added 2 commits November 10, 2024 19:41

fix readme examples

cfce3a5

approaching correctness

69bca4e

ecoskey and others added 3 commits November 11, 2024 00:25

basic sun disk rendering

06b5ba6

fix features list

cb3edf8

Merge branch 'main' into proc_sky

bbb1b45

move to cubemap

d27aa05

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Procedural atmospheric scattering #16314

Procedural atmospheric scattering #16314

ecoskey commented Nov 9, 2024 •

edited

Loading

github-actions bot commented Nov 9, 2024

pcwalton commented Nov 9, 2024

aevyrie commented Nov 9, 2024 •

edited

Loading

github-actions bot commented Nov 10, 2024

github-actions bot commented Nov 10, 2024

pcwalton commented Nov 10, 2024

pcwalton Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

pcwalton Nov 10, 2024

ecoskey Nov 10, 2024 •

edited

Loading

pcwalton Nov 10, 2024

ecoskey Nov 10, 2024 •

edited

Loading

pcwalton Nov 10, 2024

pcwalton Nov 10, 2024

ecoskey commented Nov 10, 2024 •

edited

Loading

ecoskey commented Nov 10, 2024 •

edited

Loading

github-actions bot commented Nov 10, 2024

github-actions bot commented Nov 11, 2024

aevyrie commented Nov 11, 2024

ecoskey commented Nov 12, 2024 •

edited

Loading

Procedural atmospheric scattering #16314

Are you sure you want to change the base?

Procedural atmospheric scattering #16314

Conversation

ecoskey commented Nov 9, 2024 • edited Loading

TODO

Showcase

github-actions bot commented Nov 9, 2024

pcwalton commented Nov 9, 2024

aevyrie commented Nov 9, 2024 • edited Loading

github-actions bot commented Nov 10, 2024

github-actions bot commented Nov 10, 2024

pcwalton commented Nov 10, 2024

pcwalton Nov 10, 2024 • edited Loading

Choose a reason for hiding this comment

ecoskey Nov 10, 2024 • edited Loading

Choose a reason for hiding this comment

ecoskey Nov 10, 2024 • edited Loading

Choose a reason for hiding this comment

pcwalton Nov 10, 2024

Choose a reason for hiding this comment

ecoskey Nov 10, 2024 • edited Loading

Choose a reason for hiding this comment

pcwalton Nov 10, 2024

Choose a reason for hiding this comment

ecoskey Nov 10, 2024 • edited Loading

Choose a reason for hiding this comment

pcwalton Nov 10, 2024

Choose a reason for hiding this comment

pcwalton Nov 10, 2024

Choose a reason for hiding this comment

ecoskey commented Nov 10, 2024 • edited Loading

ecoskey commented Nov 10, 2024 • edited Loading

github-actions bot commented Nov 10, 2024

github-actions bot commented Nov 11, 2024

aevyrie commented Nov 11, 2024

ecoskey commented Nov 12, 2024 • edited Loading

ecoskey commented Nov 9, 2024 •

edited

Loading

aevyrie commented Nov 9, 2024 •

edited

Loading

pcwalton Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

ecoskey Nov 10, 2024 •

edited

Loading

ecoskey commented Nov 10, 2024 •

edited

Loading

ecoskey commented Nov 10, 2024 •

edited

Loading

ecoskey commented Nov 12, 2024 •

edited

Loading