A world of sulphur and heat - Redesign of Kevari

 


Kevari has always been one of the hardest planets (Technically a moon) I have ever made. It's history is one of heavy redesigns, jumping between one dissatisfaction to another, never being happy with it, until now.

In the Past

The moon has always been inspired by fake-color pictures of Io, which at the time (and still do) strike me as specially pretty. A combination of colors that, even if untrue, had the potential for an incredible memorable body.

Fake color picture of Io from worlddreambank.com

Back in 2015, I was not well versed into image editing, so I used to make planets from Space Engine. Of course, having seen such picture of Io, I wanted to make something similar, so I created a color map that used the strange colors, Red, Yellow, Orange and Purple. The results, though extremely funny, were not good. It looked more like a pizza with kiwis than anything!

Fastfoward to 2017. At this year, I know how to put together a planet from height data from real life and was remaking the planets of Other Worlds. Kevari inevitably came around the corner for a new make-up. The wackyness of the previous design had left me now wanting to distance myself from the Io pictures that had started all of this and have Kevari be a more realistic depiction of the Galilean moon. Though not as crazy as the last one, the new design lacked identity. I would never be happy with it.

All three designs of Kevari


2022, Initial Struggles

The task to redesign the moon for a second time was quite the challenging one. Both previous designs had been subpar, and morale was not very high.

My first attemp was to start with some perlin noise, and with that iterate by making a main belt around the body and have some fluid simulation in blender to make it look like old "lava flows". This was a disaster. The design looked very primitive, and simulations for fluids, which could bump up the character of the moon, took a long time.

Heightmap for the first deisgn of 2022

Breakthrough, Procedurallity

At this time, I felt coding my own fluid simulator was my old chance. Now, doing this is no easy task, in fact, I know nothing of fluid dynamics, so after trial & error, I deviced my own small (and not very good) algorithm to simulate lava flows on a planetary surface. This ran on Java using the Processing Library to export any image as a .png file.

The algorithm first requires a base heightmap where the lava will flow. For this, I used some 3D perlin noise applied to the surface of a sphere. With this, a matrix is initialized for how much lava is in each pixel, and a number of random initial positions are filled with lava.

Each iteration every pixel with lava will look at its surrounding pixels that have no lava, and will decide how much lava to put in that pixel. If the amount of lava given to this neighbour is greater than lava in real life could actually fit in that spot, it gets saved as "reserve lava", which will itself be transfer to neighbour pixels in the future.



Simplified view of the algorithm

After running these steps for many iterations (In this case a thousand), I had one layer of lava-like height. I repeated this again, but having the initial height of the planet being the output of the last time I ran the algorithm. After doing this 4 times, I ended up with a heightmap that could work for a planet.

Because of how primitive this way of doing lava is, the end result had artifacts, seams. For this, I manually fixed them. Even with this, the result was amazing. Never this moon has had anything to go for, and this was the chance!

With this, I added different features, like craters at the equator of the moon (As it sits between the rings of an Ice Giant) and Pits (Like real life Io)

End Heightmap of the Algorithm, still with some artifacts

With the elevation data finished, I had to make the color map. This was scary, very. I wanted to go back to the initial idea of saturated and exaggerated colors, but last time I had done that, I ended up with a pizza, not a planet.

Though the task was daunting, I persisted.

An idea struck. For each of the 1000-long iterations I did for the height, I could color those regions of lava in different ways. One could be white, another purple, one subtle green.

Against all odds, after some tweaking, I had a color map that I actually liked. What I didn't expect in 2017 happened, and a good Kevari had been born.

With the main color achieved, I darkened the equatorial craters, volcanic pits and added huge recent volcanic debris like those seen at Io.

In order: A) Recent Volcanic Debris; B) Equatorial Crater; C) Purple/White Belt; D) Green Regions

Final Color Map

Conclusion

I can not express how happy I'm with the end product. After so many years, I have a volcanic planet that I'm happy with.

For the future, I would have to make a map of its biomes, which might not be so hard as I have a lot of features of the body separated in different layers.

Will I use coding/custom procedurality for making planets in the future? Probably actually. It worked (almost) flawlesly this time, so if it was help, I would surely use it again.


Poster

Back in 2017 I made posters for every planet and moon of Other Worlds. Kevari's was very dull, as I wasn't happy with how the moon looked, but this time, I am proud of it.


Comments

Popular posts from this blog

A Lava planet - Redesign of Troni