Open nvkelso opened 8 years ago
The scaling factor used to normalise the normals needs to be fixed for mercator projection. The reason it's there in the first place is so that the halving of the x
and y
directions as the user zooms out don't cause the peaks to "exaggerate", but it looks like we need a further adjustment so that peaks in high/low latitudes don't get spread out in the x
direction or exaggerated in the y
direction.
The scaling of heights needs to be refined.
Because the tile spans so much north-south, some cos(lat) per cell might be useful?