For a couple of months now, I’ve been involved in the work of deliberately creating some pretty Bad Fractals, as I call the project. Here, the goal, which is turning out swimmingly, is to find those “10000” solutions that won’t work,” to find and better by far those solutions, techniques, tricks, and tips that do to make better images. Here I have three images, two Bad, progressively less so, before reaching the third, and final image of the set. So, I’ll comment on each, and discuss what I like and don’t like about it, and why I think the Bad Fractals are just that. The images below are all courtesy of Fractal Domains for Mac. Let’s start with the first, and the worst of these! This post’s images use a modified Newton’s Method fractal type, and I’ve added terms to the formula’s numerator with higher exponent values. Bad Fractals in this series use 800 x 450 pixel resolution unless otherwise indicated:
I have serious problems with this one. While the pixelation with this could be corrected, the real issue here is the color maps used, both in the orbit traps and the fixed points. In short, this one’s colors are far too busy. In the orbit traps they bleed into each other and spoil the image’s detail, blurring the definition of the orbit traps and so spoiling the esthetics of the piece. The fixed points outside the orbit traps are too garish for my tastes. I attempt to fix those issues with the image just below, but even this has problems . . . .
This one is a bit worrisome too, even with the color maps adjusted. One, the color map of the fixed points clashes with that of the orbit traps, and two, the orbit traps are split by the orbits of the fractal set, which means I have to define the color map for each orbit region. I have improved things a bit here by lowering the range of colors available for color randomization, my usual method before adjusting color maps manually, so what appears immediately above isn’t that busy. Still, even with a reduced range of colors involved, the mismatch between the earthy shades of the fixed points and the more floral-looking orbit traps becomes apparent. So let’s try for something that doesn’t clash so much:
This one is a reasonably full-sized 1600 x 900 wallpaper, so the pixelation isn’t nearly as problematic. I’ve finally got the image to something a bit less Bad, just a wee bit, because I fixed the color map problems and the orbit trap settings of the previous two. This one had a maximum color map selection by use of randomizing somewhere between 9 and 12, and a minimum of 2, and this greatly improved the colors of the first. This one is much better even than the second in that there was no region-splitting for the orbit traps with the colors I got, so those don’t look out of place with the maps for the fixed points. The colors for traps and points fit much better together here, as I took more time selecting a better match for each. I think that this one seems a better image as a result, but you be the judge using your own sense of taste, not mine!
Tf. Tk. Tts.
The images shown in this post are, unless otherwise stated, the personal creations of the author, and unless specified to the contrary, are copyright ©2017 by Troy David Loy, AKA Troythulu. You are free to use the images here in any non-commercial way on your own venue(s), as long as you credit the source or post a link to that source online. Thank You!