[glTF] incorrect Model.boundingSphere computation?

ver. b29

Hello,

Seems to me Model.boundingSphere computation is incorrect. For example it can be seen when all model’s meshes (vertices) are moved far away from the coordinate origin in one direction.

Just in case I attch example file. It is SandCastle CesiumAir model moved along an axis by 300 units (model’s length is about 8 units). Done with SketchUp.

When SandCastle “3D Model” example runs with this model, camera view/behaviour and parameters of boundingSphee become strange.

I think that the situation when vertices of a model are not grouped around the coord. origin is not rare in practice.

Thank you,

Alexei

300.zip (757 KB)

addition

When debugShowBoundingVolume = true “debug sphere” is shown correctly

Thanks for the info, Alexei. I submitted #1810 for us to take a look. It will most likely be after Cesium 1.0.

Patrick

Alexei - do you happen to have your edited .dae for us to convert to the latest version of Cesium? We may have a fix for this.

Patrick