You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
#10384 added correct 2D / CV projection to ModelExperimental, but it doesn't properly handle models that cross the IDL. For example, a green unshaded cube will be stretched across the entire map:
These artifacts also appear in the projected S2 globe:
@ptrgags made a diagram of what was likely happening. Due to the IDL and wrapping of coordinates, one side of the model is rendered 360 degrees off from where it should be.
Unfortunately there is no way to adjust the longitude / latitude of the positions to modulo them before projecting; it's not possible to derive them from the model matrix. The most thorough solution might be to follow what GeometryPipeline does: split the geometry along the IDL, and render them both.
The text was updated successfully, but these errors were encountered:
#10384 added correct 2D / CV projection to
ModelExperimental
, but it doesn't properly handle models that cross the IDL. For example, a green unshaded cube will be stretched across the entire map:These artifacts also appear in the projected S2 globe:
@ptrgags made a diagram of what was likely happening. Due to the IDL and wrapping of coordinates, one side of the model is rendered 360 degrees off from where it should be.
Unfortunately there is no way to adjust the longitude / latitude of the positions to modulo them before projecting; it's not possible to derive them from the model matrix. The most thorough solution might be to follow what
GeometryPipeline
does: split the geometry along the IDL, and render them both.The text was updated successfully, but these errors were encountered: