Hello,
First of all, thank you for your work on Cesium and for taking the time to look into my question.
I am currently evaluating a point cloud workflow and have encountered an issue that I hope you can help me understand.
I uploaded a point cloud in LAZ format to Cesium ion and used the Cesium ion conversion workflow to generate 3D Tiles.
After the conversion, I noticed a difference between the original point cloud and the one displayed in the Cesium JS viewer.
In the original point cloud, the points are arranged along oblique scan lines, whereas in the Cesium JS viewer the points seem to be distributed on an orthogonal grid.
Would you have an explanation for this behavior?
Because of this, I am losing the level of resolution I need to properly use the point cloud.
I should also mention that the dataset is originally referenced in EPSG:3944, and the Cesium ion workflow converts it to a global coordinate system as part of the 3D Tiles generation process. Could this reprojection, or any other step in the Cesium ion conversion pipeline, be responsible for the apparent loss of resolution and the differences observed between the original point cloud and the resulting 3D Tiles dataset?
To make the comparison easier, I have included two screenshots below :
LEFT – Original LAZ point cloud displayed in Potree
RIGHT – Same dataset converted with Cesium ion and displayed in the Cesium viewer
As shown in the screenshots, the point pattern and perceived density differ significantly between the two visualizations, even though they originate from the same source dataset.
If anyone has seen this before or can point me in the right direction, I’d be very grateful 
.
Thank you in advance for any insight you can provide.
Best regards,
Elise