Here is what is coming next in TerreSculptor 2.0.
The next build of TerreSculptor will probably take a bit longer than the usual time between releases.
What I am working on right now is a rewrite to about 35% of the file load and file save methods that manage binary and text heightmap files.
This includes the Digital Elevation Model file formats of ESRI ASC, USGS BIL, GridFloat FLT, SRTM HGT, and GeoTIFF.
The first purpose of this rewrite and update is to support Void Fill on these DEM formats. Many online repositories of DEM data contain files that include voids in the dataset. The Void Fill functions in TerreSculptor provide multiple methods for dealing with these voids of missing altitude data.
The second purpose of this rewrite is to support all DEM formats on the Tile Import dialog. The Tile Import dialog additional file format support is an important feature to have as it allows for proper loading of Digital Elevation Model tiles. For larger terrain sizes, including UE4 and UE5 World Composition, having support for massive DEM terrains sourced from DEM tiles is a must.
On a related side note, one of the upcoming features for TerreSculptor is Big Array support. This will allow for better massive size terrain support limited only by computer system memory.
Currently most computers are limited in terrain size due to memory fragmentation which often causes an allocation fail at larger than 24576×24576.
Big Array will get around this fragmentation issue and allow for using large chunks of non-contiguous system memory.
This does not mean that a person can necessarily create vastly massive terrains as there are still limits in most video game engines.
UE4, and I would expect UE5, is still limited to an 8129×8129 Landscape.
TerreSculptor’s default current maximum terrain is 65536×65536.
For examples of massive terrains, a 131072×131072 World Composition in UE4/UE5 will require 34GB in 16-bit heightmap data. And a 262144×262144 World Composition in UE4/UE5 will require 137GB in 16-bit heightmap data.
Most current typical computer systems simply don’t have the memory for terrains of that size.