Right now the code normalizes the tiles by putting them all inside a top level container. So the individual glTF tiles may have a translation of X, Y, Z, and the top level container has negative X,Y,Z.
it would be cleaner to process the geometry itself so it is translated/scaled without having a hierarchy that just reverses the translation. This would make it easier to work with in Blender or build other transformations/manipulations at runtime
Right now the code normalizes the tiles by putting them all inside a top level container. So the individual glTF tiles may have a translation of X, Y, Z, and the top level container has negative X,Y,Z.
This happens here: https://github.com/OmarShehata/google-earth-as-gltf/blob/3ca254839345b83b0a707e01275b3d20490ea046/src/Viewer.js#L73
it would be cleaner to process the geometry itself so it is translated/scaled without having a hierarchy that just reverses the translation. This would make it easier to work with in Blender or build other transformations/manipulations at runtime