Volumetric-rendering in avogadro project.

[perminder-17]

Some results which are good.

When the alpha is low.

simplescreenrecorder-2025-03-14_23.50.15.mp4

When the alpha is high.

simplescreenrecorder-2025-03-14_23.46.00.mp4

You can see I have changed the loop order in the cube, which is still not perfect and will break the flying-edges code so we need to find any workaround for it. Which means (different loop order for flying edges and different loop-order for volumetric rendering).

With original flying-edges loop order for cube data (BENZENE)

With updated loop order for cube data (BENZENE)

SOME TODO'S

1. Resizing the framebuffers when we resize the screen.
2. Updating the cube's data with correct loop order
3. Scaling the volumetric mesh compared with the actual molecule rendered on the screen.
4. Adding colors support from UI.
5. Adding different sets of colors texture (2 colors, one for positive and one for negetive, @ghutchis will help in this. :) ).
6. Cleanup in the code and removing unnecessary code blocks.

[perminder-17]

I tried fixing the looping order of the cube's data, still some work is left but here's the result. Let me know how it looks @matterhorn103 @ghutchis @cryos

Water (once we get different color maps, it will look more better (h2o-dense.cube)

Screencast.from.16-03-25.10.16.23.PM.IST.webm

Ammonia

Screencast.from.16-03-25.10.18.28.PM.IST.webm

crh206

Also, for water if you use h2o-elf.cube

You get Noel's version. here(https://baoilleach.blogspot.com/2011/08/electron-density-is-not-isosurface.html)

If color mappings are different, you can get the exact same appearance.

[ghutchis]

Here's the compile issue.
cube.patch

[ghutchis]

I'm running into some crashes on Mac with the VAO in the volume render. I think it's Mac-specific, so I'll do some reading / debugging.

  // Box geometry for front/back pass
  glGenVertexArrays(1, &d->volumeBoxVao);
  glBindVertexArray(d->volumeBoxVao);

The first line crashes. I'll investigate further.

[perminder-17]

No problem geoff, I can also help. But, can you undo all the changes at cube.cpp file. (Like replacing cube.cpp file from actual main branch of cube.cpp file.) This will change the looping order and will crash some of the molecules, but you could still work on small molecules. Since the compile issue is with cube.cpp file.

[ghutchis]

I mean, you can grab the cube.cpp file yourself: https://raw.githubusercontent.com/OpenChemistry/avogadrolibs/refs/heads/master/avogadro/core/cube.cpp

Or you can git checkout master then copy the cube.cpp file to another directory, switch back to your branch and move the cube.cpp file, etc.

(There are other ways to do this in git, but these are easy.)

[perminder-17]

mean, you can grab the cube.cpp file yourself: https://raw.githubusercontent.com/OpenChemistry/avogadrolibs/refs/heads/master/avogadro/core/cube.cpp

Oops, I thought this will fix the compilation issue in your mac, but I think it's not related to the change I did in cube.cpp file.

Screencast.from.23-03-25.04.26.00.AM.IST.webm

This is the colorMapping I am trying @ghutchis @matterhorn103 @cryos

[perminder-17]

Screencast.from.23-03-25.05.20.25.AM.IST.webm

Here's the structure for caffeine from caffeine.cube file.

[perminder-17]

Oops, I forgot to change the version. I'll fix this very soon, but @ghutchis do you think this could be the reason of compilation issue in macOS? Also, if not, I think we could include some more libraries as we did in other files
#include <GL/glew.h> or maybe more to fix this issue?

[perminder-17]

Btw, @ghutchis I have lower the version in shaders. Can you take a look and also, I have link the colors (here's the result)

Screencast.from.23-03-25.09.44.34.AM.IST.webm

I could probably add a resolution as well as opacity meter in the future. Low resolution ( high fps) , High resolution (low fps).
I think for opacity meter, @ghutchis do you think it's necessary?

[ghutchis]

I don't think it's needed to go back to an old shader version. The problem on the Mac isn't the shader version, it's that glGenVertexArrays, etc. are only available in 3.x or newer and Apple only provides 2.x in compatibility mode or 4.x in core profile.

So I'll set this up to only build for Windows and Linux right now, and push to update the remaining bits of code to use core profile.

[avo-bot]

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

https://discuss.avogadro.cc/t/volumetric-rendering-progress/6865/1

[ghutchis]

You are using some really old bits of code. Please replace this:
https://gist.github.com/ghutchis/41ab922e800a7c2f14a2cb64ad65cd20

[ghutchis]

Just a minor thing - it would be great to fix the qtplugins/CMakeLists.txt - you're using an old version for some reason...

[ghutchis]

I'd probably leave all these shaders as 330 core

[ghutchis]

I'm trying on an Ubuntu VM - it seems when I resize the window, the volumetric cloud gets misaligned from the molecule:

[perminder-17]

Screencast.from.28-03-25.06.20.00.AM.IST.webm

I think we were already having the correct width and height in the volumeGeometry class :"). I was thinking very complex logic to find it out but had an oversight that we already have them.

[perminder-17]

Hi @ghutchis , do you mind checking it once again then I could probably work on code cleanups.

[matterhorn103]

I'm trying on an Ubuntu VM - it seems when I resize the window, the volumetric cloud gets misaligned from the molecule:

I believe this is actually part of an existing problem and applies to all displayed orbitals, it's not isolated to the volumetric renderer. Does it also happen for you with isosurfaces?

I noticed it a little while ago and just hadn't got round to making an issue for it yet.

(I figured there's enough more urgent issues that I didn't need to rush with the bug report.)