Implement Premultiplied Alpha for OIT

[Schmarni-Dev]

Objective

• Allow rendering objects with AlphaMode::Premultiplied and AlphaMode::Add while using Order Independent Transparency.
• Fixes Rendering Ordering Issue with OIT and AlphaMode::Premultiplied #20373.

Solution

• Store Premultiplied color in the OIT buffer and premultiply in the case of AlphaMode::Blend, this shouldn't have any downsides since blending was done with premultiplied alpha anyway.
• It might be useful to actively break the oit_draw signature by adding a premultiply bool to ensure that shader authors are aware of the changed meaning of the color argument.

Testing

• These changes have been tested by running the modified order_independent_transparency example, an older version of these changes has been in use in a project i work on.

---

Showcase

Here the Red Sphere uses AlphaMode::Blend, the Blue Sphere AlphaMode::Add and the Green Sphere AlphaMode::Premultiplied.

[github-actions]

Welcome, new contributor!

Please make sure you've read our contributing guide and we look forward to reviewing your pull request shortly ✨

[IceSentry]

Thank you for making this PR! We are about to merge a big OIT refactor though so you will likely have some conflict with your PR #21831

[Schmarni-Dev]

I can rebase these changes once that PR is merged

[beicause]

AlphaMode::Add operation is order independent so it doesn't need oit_draw.
😅 AlphaMode::Add seems to be result = vec4(src_alpha * src_color.rgb, 0.0) + 1 * dst_color.rgba which is not order independent.

It seems multiply is the only unsupported mode as it requires a different blend function, which is not free.

[beicause]

Can we do this somewhere before premultiply_alpha?
For Add and Premultiplied modes we should not cull alpha in here

[beicause]

Or we could consider removing this setting. After all users can do it themselves in custom materials.

[Schmarni-Dev]

this should be fixed now, i adjusted the check to work correctly for premultiplied alpha, should have the same results as before this change

[beicause]

It is not correct to compare individual components. An alternative is to compare luminance.

I think there should be a luminance threshold setting for premultiplied or add mode.

[Schmarni-Dev]

why would it not be correct to compare the individual components? it should be the same check as the one that was here before, just now it acts on premultiplied color

[beicause]

Considering the color contribution of premultiplied alpha, it seems more reasonable to compare the emitted luminance rather than comparing each component individually