Open PrimaryFantasty opened 3 weeks ago
(Just a cross-link to the forum post: https://community.cesium.com/t/why-modified-the-behavior-of-khr-texture-transform-extension-since-cesiumjs-1-97/32984 )
The fract is incorrect.
Thanks @scurest , I also suspected that. I'm not entirely sure how to confirm that to an extent that would allow closing this issue. Using https://github.com/cx20/gltf-test to look at the TextureTransformTest
, namey rendered in CesiumJS, it looks correct:
With the small caveat that it's not clear whether the model contains texture coordinates where fract
would actually make a difference. (I suspect such models are rare to begin with).
I'd vote to close this, saying that the fract
does not seem to make sense (and is also not used in the sample GLSL code of the specification). So I assume that this was a bug in earlier CesiumJS version, but has been fixed with the "Model" rewrite that happened in 1.97.
The fract is incorrect.分形不正确。
Thanks for your reply!
Thanks @scurest , I also suspected that. I'm not entirely sure how to confirm that to an extent that would allow closing this issue. Using https://github.com/cx20/gltf-test to look at the
TextureTransformTest
, namey rendered in CesiumJS, it looks correct:
With the small caveat that it's not clear whether the model contains texture coordinates where
fract
would actually make a difference. (I suspect such models are rare to begin with).I'd vote to close this, saying that the
fract
does not seem to make sense (and is also not used in the sample GLSL code of the specification). So I assume that this was a bug in earlier CesiumJS version, but has been fixed with the "Model" rewrite that happened in 1.97.
Thank you! In fact, The first time I found it was that in Esri I3S uvRegion, which contains the concept of uvRegion. Certainly, It's different from"KHR_texture_transorm", I tried converting it into glTF by adding a custom glTF extesion in Cesium to adapt "uvRegion", but the same to you, I'm not sure the meaning of this detail, cause of CesiumJS 1.96, 1.97 and "uvRegion" in I3S.
@javagl I made a typo, sry~
Current:
But in CesiumJS 1.96:
The difference between them is that interpolated UVs has been "fract" before transforming,and they make different effect. So what kind of them should I take?