I wrote a script in the Processing language which randomly generates colored, nested circles on a grid akin to my cousin Daniel Bartholomew's work of the same title. When the Processing script runs, it animates the circles, and if you tap on them, their color animates. I entered it in the Springville Museum of Art's 34th Spiritual and Religious Art of Utah Contest (if it makes it into the show, it will be displayed on a large kiosk). [2019-10-04 UPDATE: This work made it into the show! It was on display at the Springville Museum of Art, October 16, 2019 – January 15, 2020.] Here is the artist statement:
"..by small and simple things are great things brought to pass.." -Alma 37:6
Tap or swipe circles and watch what happens!
Just like your interaction changes this work, I believe that God interferes with reality–sometimes to dazzling effect. I believe that mere existence is amazing besides, or if not, filled with promise.
Images you interact with are "tweeted" @earthbound19bot (Twitter social media).
I coded this in the Processing language with Daniel Bartholomew's support and input. It imitates his original pen and marker works of the same title, with animation, and generating any of about 4.3 billion possible variations at intervals.
BSaST v0.9.13 seed 1713832960 frame 133
I dedicate all these images to the Public Domain. I can literally make 4.3 billion other ones if anyone "steals" these. [UPDATE 2: The kiosk saved as many user-generated works from interactions with it as it could, and I've archived them in my "firehose" gallery here.]
[UPDATE: there's a lot more to light and color science than I perhaps inaccurately get at in this post. Also, try color transforms and comparisons (if the latter is possible?) in Oklab.]
I wrote a Python script that uses that library to sort any list of RGB colors (expressed in hex) so that every color has the colors most similar to it next to it. (Figuring out an algorithm that does this broke my brain–I guess in a good way.) (I also wrote a bash script that runs it against all .hexplt files (a palette file format which is one RGB hex color per line) in a directory.)
The results are better than any other color sorting I've found, possibly better than what very perceptive humans could accomplish with complicated arrays of color.
Here's an image of Prismacolor marker colors, in the order that results from sorting by this script (the order is left to right, top to bottom) :
For before/after comparison, here is an image from the same palette, but randomly sorted; the script can turn this ordering of the palette into the above much more contiguous-appearing:
(It's astonishing, but it seems like any color in that palette looks good with any other color in it, despite that the palette comprises every basic hue, and many grays and some browns. They know what they are doing at Prismacolor. I got this palette from my cousin Daniel Bartholomew, who uses those colors in his art, which you may see over here and here.)
Here is another before and after comparison of 250 randomly generated RGB colors sorted by this script. You might correctly guess from this that random color generation in the RGB space often produces garish color arrays. I wonder whether random color generation somehow done in a model more aligned with human perception (like CIECAM02) would produce more pleasing results.
See how it has impressive runs of colors very near each other, including by tint or shade, and good compromises when colors aren't near, with colors that are perceptually further from everything at the end. Also notice that darker and lighter shades of the same hue tend to go in separate lighter/darker runs–with colors that well interpolate into those runs in between!–instead of having lights and darks in the same run, where the higher difference of tint/shade would introduce a discontiguous aspect.
Tangent: in RGB space, I tested a theory that a collection of colors which add (or subtract!) to gray will generally be a pleasing combination of colors–and found this to be often true. I would like to test this theory in the CIECAM02 color space. I'd also like to test the theory that colors randomly generated in the CIECAM02 space will generally be more pleasing alone and together (regardless of whether they were conceived as combining to form gray).
I really can't have those as the last images in this post. Here is a favorite palette.
[Edit 2020-10-07: I had renamed or moved several things I linked to from this post, which broke links. I corrected the links after a reader kindly requested to know where things had gone.]
This is one of thousands of images like it (each unique though) I've recently generated with an experimental process. The experiment is a success if I may say so.
Phase I.
– collect several cool satellite images of civilization and/or wilderness, e.g. from this site: https://earthview.withgoogle.com/
– for every image pair in the collection, make a "diff" image (subtract the RGB values of every pixel in one image from every pixel in the other image), and save the result
– for every image pair in the collection, make an averaged image (average the RGB values of every pixel in one image with another), and save the result
Phase II.
– liberally delete less impressive results
Phase III.
– for every diffed result, average it with an averaged result and save that.
– for every averaged result, subtract (diff) a diffed result.
– liberally delete less impressive results. Good luck–with 17 source images and heavy pruning in Phase II, this will give me 17k+ results, so far all of them compellingly cool.
(Phase IV: sort all results by approx. nearest most similar and string them together in a movie of crossfades to see works between the works.)
(Phase V: accidentally produce glitch art because your computer ran out of hard drive space and memory doing all this, but the processing script keeps calling the utilities that do this, and the utilities break. I'll post some glitch results later).
(Phase VI: realize you have a storage and bandwidth problem for your new many gigabytes of images.)
Base work created by Filter Forge auto-collage filter custom setting. I probably also used a custom variant of the SideToSide filter; that in built up alpha and hue layer variations to produce rectangular hue/tone variety. I might like to call this post-plasticism (after Piet Mondrian's neoplasticism; this is structurally similar but uses any color).
This first is vector art (an svg), which you may save and reuse. You may reuse these works freely under Creative Commons Attribution 4. I'd appreciate credit in reuse.
The animated variant is concieved as unobtrusive decorative video art. Or maybe it would be distracting. I don't know, because I don't know who displays art as such. Do you?
The following variant and resource images which I made along the way, I release into the Public Domain:
Variant via the Filter Forge "side to side" filter by Skybase:
An alpha resource via the Filter Forge Terrain Hightfield Generator by LigH; I used this (and variants of it) as a transparency channel in filter layers to make uneven interesting application of filters:
Here's a presentation of choice output from the batch in the previous post. Here also is a link to the vector sources, which I release into the Public Domain.
I made black and white vector art of a doodle and did random color fills of the blank areas from a palette. I automated this with bash scripts and CLI utilities on Windows+MSYS2.
I also strung many of these images together in an animation mostly sorted by next most similar image (also with scripts).
Thumbnails link to larger resolution images, free for personal use.
I think it's a happy coincidence that my first run of the random color fill script produced grass-like color below and sky-like color above.
I've changed the name of this since after I gave up numbering works (cataloguing them is a technical feat that got away from me), and starting naming some works that I have no better name for by date. This was originally named and titled Work 00091 [+some description about abstract line art with random color fills]. Later, in syndicated media posts I renamed it 2016-11-24, which date is an error. My files named by date for this are 2017-01-13 (with redevelopment later in 2021). Blahuarg.
Click or tap the image for a high resolution version, free for personal use. If you want a print please leave a comment. ~ Based on something I started way back in 2003. Gah. I'm old. This is however a near overhaul of that. Created with Corel Painter, Filter Forge, and Photoshop. ~ A hoity-toity robot talks about this at: http://s.earthbound.io/artgib
Works utilize this medium today. A concrete form with various organisations, including the work /Black Out/, in which imagination, dreams, and death are largely intuitive: preferring that the settlers went on to study sociocultural trends in photography, media and intellectual creativity in masterpieces of probably the most loved French post-impressionist masters.
Produced via http://s.earthbound.io/autobrood ~ I randomly generated a lot of fractals, picked pretty ones, and interbred them. This fractal results from alternating the genes of 2016 11 14 21 22 46 962407600 and 2016 11 14 21 53 58 616402600.