-
Notifications
You must be signed in to change notification settings - Fork 3.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
glTF 1.0.1 Adds support for normalized vertex arrays and a test #4247
glTF 1.0.1 Adds support for normalized vertex arrays and a test #4247
Conversation
@pjcozzi, could you look at this if you get a chance so we can start getting this chain of pull requests merged? |
@lasalvavida I should have thought of this sooner, but I created a gltf-1.0.1 branch which we can use to accumulate all glTF 1.0.1 updates, and then we'll merge to master in one go once the spec is official. I also updated CHANGES.md in this branch so each PR does not need to provide fine-grained updates. Can you please change this and all glTF PRs to target the gltf-1.0.1 branch? This is a new option in the GitHub UI. |
Changed the base to the gltf 1.0.1 branch |
@lexaknyazev we are starting to implement glTF 1.0.1 in Cesium and gltf-pipeline. This should help tighten up anything that we overlooked. It's totally optional, but if you have the time, it would be great if you could look over these PRs to make sure we are keeping the spec/implementation in-sync. We'll CC you on them; each PR is small. |
@@ -1598,6 +1599,13 @@ defineSuite([ | |||
}); | |||
}); | |||
|
|||
it('loads a glTF with WEB3D_quantized_attributes and accessor.normalized for higher precision decode matrix', function() { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not following the "higher precision decode matrix." Why not add per-vertex normalized ubyte8
colors to an existing model since that is the common use case? In general, I would also like to avoid bloating the repo with a lot of new test models to test corner cases if we are able to augment existing ones reasonably.
e09a4bf
to
a8afa5e
Compare
This has been updated to modify the existing quantized box model instead of adding a new one. Commits have been squashed so it is removed from the history. That may be a more common use case, but this also tests the functionality of See my gltf-pipeline comment on why this is useful. |
This looks good to me. I think @pjcozzi's suggestion is good since the testing can be more concrete, i.e. checking the pixel color vs. just checking it was rendered. It should be pretty easy to modify Edit: I think it's fine to keep the current test too. |
Modified Box-Color to use the normalized property |
Looks good! @pjcozzi do you want to do a quick final review? |
I see, thanks! |
FYI - this does not update the |
Part of #4009. Test model is quantized box with a high precision decode matrix via the glTF 1.0.1 normalized property.