WIP Topographic Map and Question Regarding Fractal Surfaces in Wilbur/FT

[whitemiasma]

I've been working on trying to get my topographic map for my world to be as realistic as my abilities will allow me to. I've gotten to a point where I'm happy with how my mountain ranges and river beds are placed, but I feel like overall the image may be too smooth. For example, when I start to create regional maps there may be very little topographic variability in them. I feel like a good way to combat this is by fractalizing the surface, but I wish to do this without completely altering the terrain of the world. I've tried fiddling with adding fractal noise in Wilbur but haven't been able to produce anything I'm happy with. Does anyone have any advice on how I could go about creating some cool topographic features without replacing my existing ones? Preferably using Wilbur or FT. My WIP topographic map is below.

Topo map low res.png

[waldronate]

I don't know if it's important to your map, but you have a whole lot of surface at very low altitude and then a lot at middle-range. I will assume for the moment that the low-lying land (under 20) doesn't matter and that the grid is a decoration rather than an important part of your altitude map.

The simplest way is to just do a little more of what you've likely done already in Wilbur:
* select the land areas
* add a bit of noise (roughly 10% of your overall altitude range)
* fill basins
* incise flow (amount=2, exponent=0.3, effect blend=1)
* precipiton erosion, 2 passes
* fill basins
* incise flow (amount=1, exponent=0.6, effect blend=0.2)
* precipiton erosion, 1 passes
* fill basins

That should crinkle things up a little without wiping out your existing terrain too much.

[whitemiasma]

Thanks Waldronate,

I've gone through cycles of adding 10% noise and running 25 passes of precipiton erosion a few times in Wilbur. I will give these steps a try when I get a chance. I am concerned about the amount of low lying land, I would like to have more highlands and mountain ranges to be present but would like them to be more random if possible. Do you think that selecting all the low lying areas and calculating a height map for them would be a good idea? Then I could run the steps you provided to blend my existing mountains back in. Not sure if that's a good approach, might give it a shot...

[waldronate]

25 passes of precipiton erosion will make things very smooth. It's hard to be sure from your heightmap exactly which parts are intended to be water and which are intended to be land.

The low-lying areas look like you didn't have a selection during a run of precipiton erosion and they may not have been intended to be there.

[whitemiasma]

25 passes of precipiton erosion will make things very smooth. It's hard to be sure from your heightmap exactly which parts are intended to be water and which are intended to be land.

The low-lying areas look like you didn't have a selection during a run of precipiton erosion and they may not have been intended to be there.

The black areas are intended to be water, with all grey and white areas being land. The inland black areas are intended to be seas and lakes, are those the low-lying areas you're referring to? Maybe I should have excluded inland water bodies from my land-sea mask and added them in afterwards?

[waldronate]

If you load that PNG map that you posted into Wilbur as a height map, you will see that there are large areas of near-black things above the main continent. Those are the low-lying areas that I was referring to.

[whitemiasma]

Just tested and I see what you mean now. I have this issue with reloading images that I've already worked with in Wilbur, they get raised ~ 30m somehow. If you do a mathematical offset of the whole map of -30 it should appear how it is intended to.

I've attempted to fix that in this image below. I also tried the steps you suggested and I'm thinking it looks pretty good! Would be curious to know what you think. Thanks for you help.

Topo map low res_v2.png

[waldronate]

A PNG file only stores unsigned (0 to 255 for 8-bit or 0 to 65535 for 16-bit) data. The original Wilbur data is likely negative below sea level, so Wilbur assigns the lowest value on the surface to 0 in the PNG and the highest value on the surface to 255 or 65535 in the PNG. This offset is what you observe.

That result looks about like what I would expect for the process that I recommended.

[johnvanvliet]

i use the BT format "binary Terran" it uses 32 bit floating point data

i like to use a "emboss" filter for looking at changes in 16 bit signed and unsigned and 32 bit floating point images

an example
a fractal output of one of the generators
-- 8 bit copy of the 32 bit float



you really can not tell to much from that , but a shaded image is a bit different


the second one is after 9 cycles
3 erosion cycles a 1% for overall
and 6 erosion cycles a 1%for "per pixel"

i started with a signed tiff with pixel values from ( -216 to a +6654 ) converted to 32 bit float
gdal and gmic works great for converting

convert 16 s to float

gmic input16bit.Image.tiff -o 2k.tiff

gdal_translate -of bt  2k.tiff 2k.bt

open in wilbur
run
convert back to a 32 bit float image

gdal_translate  3.bt 3.tiff