Is there any advantage using Real-world map data over World machine generated height map?

[jake_lee]

As the title said, is there any advantage using real world data over WM generated heightmap? Does it create a more realistic feeling? To me WM generated map looks quite convincing already…
My current workflow is referencing from real world photos, but generating and using all the heightmaps from WM. What should I be cautious of?

[HYLK]

Realworld data looks, by definition, realistic. The big caveat is the quality/resolution of the data. Another disadvantage is the lack of control and lack of data for erosion and such. If you do everything in WM, you get more “extra” data, however, to get proper realistic terrain costs a lot of time and experience.

The middleground is to import realworld data and to slightly alter it in WM, which can give you back erosion data and such, at the loss of some alteration.

Another benefit of doing everything in , is artistic freedom. You can decide where to place what, where to follow the rules of nature and where not etc.

[Stephen]

As Hylk said, the problem with real world data is mostly in the resolution of it.

I like to think of terrains existing at three scales. This is not just a conceptual aid; different world scales have different dominant processes.

1. Continent / Planetary. Dominant processes: Plate tectonics, global climate

2. Regional. Dominant processes: Water flow, hillslope evolution

3. Human-scale. Dominant processes: Rock fracture, Plants

Looking at the above, you can see that World Machine clearly targets regional scale. Continental scale is also supported but not as a primary focus (there is not, yet, a plate tectonics device ). At the other extreme: human scale, heightfields are not a good match for actual terrain.

Most of our real world terrain data is squarely in category 1 due to its low resolution. So one very common use of World Machine is taking “category 1” real world data and adding detail to bring it to category 2.