Add depth-of-field (Blurring rendering options) #1662

perminder-17 · 2024-04-11T02:31:04Z

issue : #347
Signed-off by : permindersingh089@gmail.com
Developer Certificate of Origin
Version 1.1

Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
1 Letterman Drive
Suite D4700
San Francisco, CA, 94129

Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.

Developer's Certificate of Origin 1.1

By making a contribution to this project, I certify that:

(a) The contribution was created in whole or in part by me and I
have the right to submit it under the open source license
indicated in the file; or

(b) The contribution is based upon previous work that, to the best
of my knowledge, is covered under an appropriate open source
license and I have the right under that license to submit that
work with modifications, whether created in whole or in part
by me, under the same open source license (unless I am
permitted to submit under a different license), as indicated
in the file; or

(c) The contribution was provided directly to me by some other
person who certified (a), (b) or (c) and I have not modified
it.

(d) I understand and agree that this project and the contribution
are public and that a record of the contribution (including all
personal information I submit with it, including my sign-off) is
maintained indefinitely and may be redistributed consistent with
this project or the open source license(s) involved.

Signed-off-by: Perminder <permindersingh089@gmail.com>

perminder-17 · 2024-04-11T03:03:00Z

Hi @ghutchis code is functioning correctly, but there are a couple of concerns I'd like your help with. Also I am thinking to add visual tests when the all the requirements will be done. Here's the other PR OpenChemistry/avogadroapp#492

The code currently uses 64 samples for blurring, following the number of samples from Speck. However, this reduces frame rates a bit. I'm considering reducing the number of samples, but I'm concerned about compromising the quality of the blur. Do you think decreasing the samples would be advisable?
When we zoom in or out, the blur changes, which differs from the behavior in Speck. While this deviation could mimic real-world effects, such as objects appearing blurry when they're too close or too far. Considering potential future features like VR, this behavior might actually enhance realism. What's your take on this? Should we maintain or adjust the blur behavior?
Your feedback is invaluable, and I'm committed to optimizing and tidying up the code. I'd greatly appreciate your input on any minor cleanups or optimizations you think are necessary.
I was thinking about adding blur and fog effects to a new file, but I believe combining all rendering options into one would be better.
I'll include the ability to adjust the strength of blur, allowing users to control how blurry they want things to be. Will be setting some other uniforms for that.

avo-bot · 2024-04-11T03:06:58Z

This pull request has been mentioned on Avogadro Discussion. There might be relevant details there:

https://discuss.avogadro.cc/t/add-depth-of-field-fog-or-blurring-rendering-options/5332/9

Signed-off-by: Perminder <permindersingh089@gmail.com>

perminder-17 · 2024-04-11T16:35:41Z

Have a look, In the above commit I changed as you told("does not blur when camera is too close")

BEFORE (without BLUR)

AFTER-BLUR (THE INTENSITY CAN BE INCREASED)

BEFORE (without blur)

AFTER-BLUR

vinayakjeet

Review of Shader Code Enhancements

The proposed changes to the shader code, including enhanced randomness in the rand function, dynamic sample adjustment, and smoother weight transitions using smoothstep, significantly improve performance and visual quality. These updates are well-aligned with best practices for rendering optimizations.

@ghutchis , could you please review these enhancements and provide your feedback?

vinayakjeet · 2024-04-21T17:59:55Z

avogadro/rendering/solid_first_fs.glsl

+
+  float total = 1.0;
+  vec4 color = texture2D(inRGBTex, texCoord);
+  // number of samples can be optimized.


int numSamples = int(clamp(blurAmt * 64.0, 16.0, 64.0));
// dynamically reduce number of samples based on blur amount

for (int i = 0; i < numSamples; i++) {
// Remaining loop code of blur application...
}
@ghutchis your reviews for the changes proposed.

Thank you for the review. I'll definitely make time to address the suggestions. However, regarding this issue, the loop index can't be compared with a non-constant expression. Therefore, adding numSamples to the for loop won't function since its value isn't constant. Could also leads to error. Are you sure about this one??

You're right, and I appreciate your careful consideration. While GLSL 1.2 does support dynamic looping, the behavior can indeed vary based on the GPU and driver support. Sticking with your current approach to ensure compatibility and stability is wise. Thank you for pointing this out, @guthichs, and your reviews and help are greatly appreciated!

vinayakjeet · 2024-04-21T18:28:55Z

avogadro/rendering/solid_first_fs.glsl

@@ -66,6 +68,47 @@ float lerp(float a, float b, float f)
    return a + f * (b - a);
 }

+float rand(vec2 co){
+  return fract(sin(dot(co.xy ,vec2(12.9898,78.233))) * 43758.5453);
+}


float rand(vec2 co) {
return fract(sin(dot(co.xy ,vec2(12.9898,78.233)) + cos(dot(co.xy, vec2(93.9898,50.233)))) * 43758.5453);
}
//more complex pattern produced which helps in reducing the noise. @ghutchis your insights ?

vinayakjeet · 2024-04-21T18:29:00Z

avogadro/rendering/solid_first_fs.glsl

+    vec2 offset = (vec2(cos(angle), sin(angle)) * radius * 0.05 * blurAmt) / pixelScale;
+    float z = depthToZ(texture2D(inDepthTex, texCoord + offset).x);
+    float sampleBlur = calcBlur(z, pixelScale);
+    float weight = float((z >= origZ) || (sampleBlur >= blurAmt * radius + 0.));


float weight = 1.0 - smoothstep(0.0, 1.0, abs(z - origZ) / blurAmt);
// using smoothstep to soften weight transitions
This modification uses smoothstep to create a more gradual transition in weights applied to sampled colors, potentially reducing artifacts like banding. Reviews @ghutchis

I think the smoothstep() would be nice here, since it's also easier to read.

Signed-off-by: Perminder <permindersingh089@gmail.com>

perminder-17 · 2024-06-20T21:09:01Z

simplescreenrecorder-2024-06-21_01.48.58.mp4

simplescreenrecorder-2024-06-21_02.08.17.mp4

simplescreenrecorder-2024-06-21_01.57.25.1.mp4

Hi, I just have pushed the recent changes, above is the result I uploaded to show how depth-of-field is working for now. If any contributors or users of avogadro software have any suggestion, I would be more than happy to consider those changes into my code. :)

avo-bot · 2024-09-08T15:32:35Z

This pull request has been mentioned on Avogadro Discussion. There might be relevant details there:

https://discuss.avogadro.cc/t/september-live-updates/5571/1

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

github-actions · 2024-09-18T21:52:22Z

Here are the build results
Avogadro2.AppImage
Win64.exe
Artifacts will only be retained for 90 days.

ghutchis

Just some minor cleanups, but it otherwise looks good.

avogadro/rendering/solid_first_fs.glsl

ghutchis · 2024-09-20T18:27:07Z

avogadro/rendering/solid_first_fs.glsl

+    vec2 offset = (vec2(cos(angle), sin(angle)) * radius * 0.05 * blurAmt) / pixelScale;
+    float z = depthToZ(texture2D(inDepthTex, texCoord + offset).x);
+    float sampleBlur = calcBlur(z, pixelScale);
+    float weight = float((z >= origZ) || (sampleBlur >= blurAmt * radius + 0.));


I think the smoothstep() would be nice here, since it's also easier to read.

avogadro/rendering/solid_first_fs.glsl

avogadro/rendering/solidpipeline.cpp

github-actions · 2024-09-20T18:56:28Z

Here are the build results
Avogadro2.AppImage
Win64.exe
macOS.dmg
Artifacts will only be retained for 90 days.

Signed-off-by: Perminder <permindersingh089@gmail.com>

perminder-17 · 2024-09-25T23:11:13Z

I separated the fog shader into a different file to make the process faster because the Ambient Occlusion (AO) operation was running a bit slow. Thanks to @ghutchis for the helpful insights!

simplescreenrecorder-2024-09-26_04.27.26.mp4

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

ghutchis · 2024-09-26T19:16:28Z

Sorry, I guess I wasn't clear. I think the fog should be in both shaders:

AO + DOF + fog (i.e., good GPU shader)
just fog (i.e., faster rendering)

In other words, if AO or DOF or Edge detection is enabled, the "full" shader is used. If all three are off, then the faster fog-only shader is used for speed.

Does that make sense?

So my suggestion is to put back the fog into the main shader. That's the default. Then we can change which shader is used in the C++, e.g.

if (ao_enabled || dof_enabled || edge_enabled)
   // use full shader
else
   // fast fog-only shader

perminder-17 · 2024-09-26T19:23:21Z

Sorry, I guess I wasn't clear. I think the fog should be in both shaders:

Ahh..I am sorry. I thought we need to separate the fog shaders, Sorry for the misunderstanding. I will fix that now.

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

perminder-17 · 2024-09-27T01:01:18Z

Okay, I think I have worked on the suggestion @ghutchis . Let me know if you have any other suggestions?

github-actions · 2024-09-30T04:37:27Z

Here are the build results
Avogadro2.AppImage
Win64.exe
Artifacts will only be retained for 90 days.

github-actions · 2024-09-30T18:14:16Z

Here are the build results
Avogadro2.AppImage
Win64.exe
macOS.dmg
Artifacts will only be retained for 90 days.

ghutchis · 2024-10-12T00:04:21Z

I think the effect may need some tweaking (e.g., strength, position) but the code looks good. Thanks for all the work!

github-actions · 2024-10-12T01:22:54Z

Here are the build results
Avogadro2.AppImage
macOS.dmg
Win64.exe
Artifacts will only be retained for 90 days.

perminder-17 · 2024-10-12T13:40:09Z

I think the effect may need some tweaking (e.g., strength, position) but the code looks good. Thanks for all the work!

Yes, it could be improved. The problem was when you zoom-out zoom-in the change in the projection matrix is not consistent or linear. That's what makes it a problem. :"). I could probably open a new PR in improving it shortly. Also, I think we could probably merge this PR OpenChemistry/avogadroapp#492

ghutchis · 2024-10-12T14:00:43Z

It's a great effect. Exactly as you say, it's the change when zooming. Tweaking the strength and projection matrix are secondary now that the code is already merged.

I needed to confirm that OpenChemistry/avogadroapp#492 built cleanly after the merge. I'll also tweak the text a bit in the labels for avogadroapp.

perminder-17 added 2 commits April 11, 2024 07:53

Depth-blur

d527009

Signed-off-by: Perminder <permindersingh089@gmail.com>

Uniforms

d7e5b10

Signed-off-by: Perminder <permindersingh089@gmail.com>

perminder-17 mentioned this pull request Apr 11, 2024

Add depth-of-field (Blurring rendering options) OpenChemistry/avogadroapp#492

Merged

not-close-blur

f99fbb7

Signed-off-by: Perminder <permindersingh089@gmail.com>

vinayakjeet reviewed Apr 21, 2024

View reviewed changes

perminder-17 added 2 commits June 8, 2024 07:45

Strength for blur added

cd7059a

Signed-off-by: Perminder <permindersingh089@gmail.com>

dof-position added

110d754

Signed-off-by: Perminder <permindersingh089@gmail.com>

ghutchis linked an issue Sep 11, 2024 that may be closed by this pull request

Add depth-of-field fog or blurring rendering options #347

Closed

ghutchis added the rendering label Sep 11, 2024

Merge branch 'master' into depth-blur

dab63d4

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

ghutchis reviewed Sep 20, 2024

View reviewed changes

seperating shaders for speed

0c27224

Signed-off-by: Perminder <permindersingh089@gmail.com>

adding smoothstep

2f53aa4

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

perminder-17 added 4 commits September 27, 2024 06:01

cmake-file changes

2ab576c

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

applying condition to shader for seperating fog

1af7390

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

delete newly created shader for fog.

6fc981c

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

fixing solidpipeline.cpp

ef20d54

Signed-off-by: Perminder Singh <127239756+perminder-17@users.noreply.github.com>

ghutchis merged commit 818c059 into OpenChemistry:master Oct 12, 2024
19 of 21 checks passed

perminder-17 deleted the depth-blur branch October 12, 2024 14:17

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add depth-of-field (Blurring rendering options) #1662

Add depth-of-field (Blurring rendering options) #1662

perminder-17 commented Apr 11, 2024

perminder-17 commented Apr 11, 2024 •

edited

Loading

avo-bot commented Apr 11, 2024

perminder-17 commented Apr 11, 2024 •

edited

Loading

vinayakjeet left a comment

vinayakjeet Apr 21, 2024

perminder-17 Apr 21, 2024

vinayakjeet Apr 22, 2024

vinayakjeet Apr 21, 2024 •

edited

Loading

vinayakjeet Apr 21, 2024

ghutchis Sep 20, 2024

perminder-17 commented Jun 20, 2024

avo-bot commented Sep 8, 2024

github-actions bot commented Sep 18, 2024

ghutchis left a comment

ghutchis Sep 20, 2024

github-actions bot commented Sep 20, 2024

perminder-17 commented Sep 25, 2024

ghutchis commented Sep 26, 2024

perminder-17 commented Sep 26, 2024

perminder-17 commented Sep 27, 2024

github-actions bot commented Sep 30, 2024

github-actions bot commented Sep 30, 2024

ghutchis commented Oct 12, 2024

github-actions bot commented Oct 12, 2024

perminder-17 commented Oct 12, 2024 •

edited

Loading

ghutchis commented Oct 12, 2024

Add depth-of-field (Blurring rendering options) #1662

Add depth-of-field (Blurring rendering options) #1662

Conversation

perminder-17 commented Apr 11, 2024

perminder-17 commented Apr 11, 2024 • edited Loading

avo-bot commented Apr 11, 2024

perminder-17 commented Apr 11, 2024 • edited Loading

BEFORE (without BLUR)

AFTER-BLUR (THE INTENSITY CAN BE INCREASED)

BEFORE (without blur)

AFTER-BLUR

vinayakjeet left a comment

Choose a reason for hiding this comment

vinayakjeet Apr 21, 2024

Choose a reason for hiding this comment

perminder-17 Apr 21, 2024

Choose a reason for hiding this comment

vinayakjeet Apr 22, 2024

Choose a reason for hiding this comment

vinayakjeet Apr 21, 2024 • edited Loading

Choose a reason for hiding this comment

vinayakjeet Apr 21, 2024

Choose a reason for hiding this comment

ghutchis Sep 20, 2024

Choose a reason for hiding this comment

perminder-17 commented Jun 20, 2024

avo-bot commented Sep 8, 2024

github-actions bot commented Sep 18, 2024

ghutchis left a comment

Choose a reason for hiding this comment

ghutchis Sep 20, 2024

Choose a reason for hiding this comment

github-actions bot commented Sep 20, 2024

perminder-17 commented Sep 25, 2024

ghutchis commented Sep 26, 2024

perminder-17 commented Sep 26, 2024

perminder-17 commented Sep 27, 2024

github-actions bot commented Sep 30, 2024

github-actions bot commented Sep 30, 2024

ghutchis commented Oct 12, 2024

github-actions bot commented Oct 12, 2024

perminder-17 commented Oct 12, 2024 • edited Loading

ghutchis commented Oct 12, 2024

perminder-17 commented Apr 11, 2024 •

edited

Loading

perminder-17 commented Apr 11, 2024 •

edited

Loading

vinayakjeet Apr 21, 2024 •

edited

Loading

perminder-17 commented Oct 12, 2024 •

edited

Loading