With this change in place, the blobs look more interesting
May 1, 2017 3:11 PM   Subscribe

 
The process pictures while explaining the Processing seems more finished and watercolor-y than the finished faux watercolor at the finish. Pretty close to Pastels though.

Is there a way to plug this into a human controlled graphics program?
posted by tychotesla at 4:03 PM on May 1, 2017


This is nice and thought provoking. Thanks!
posted by Dmenet at 4:29 PM on May 1, 2017


I knew who posted this without even looking.
posted by Johnny Wallflower at 5:02 PM on May 1, 2017 [3 favorites]


This should be posted to Facebook's "Oh Okay Good Job Then" post. I guess I'm either a techno-Philistine or art snob, but this doesn't do anything for me.
posted by kozad at 5:14 PM on May 1, 2017


This makes me feel clever, as I've been doing something similar in Illustrator lately. Nowhere near as many translucent layers, because I'm gonna make a LOT of shapes over the course of a drawing, and they get redrawn regularly...
posted by egypturnash at 5:17 PM on May 1, 2017


I love watercolors, but I find this kinda cold and depressing, like reducing art to an computer algorithm.
posted by Pirate-Bartender-Zombie-Monkey at 7:58 PM on May 1, 2017


I was expecting to be completely unimpressed, but found the second piece he did with the technique beautiful. Neither one quite looks like watercolour, IMO, more like dyed fabric.
posted by two or three cars parked under the stars at 7:58 PM on May 1, 2017


I love watercolors, but I find this kinda cold and depressing, like reducing art to an computer algorithm.

More reducing paint to a computer algorithm, no? The placement of paint is either randomized or semi-random and partially guided by human judgement - which puts it in a similar category to some pre-software abstraction.

Is there a way to plug this into a human controlled graphics program?

Not something I'm an expert on but I'm under the impression that the big digital paint programs do contain simulations of various media and techniques.
posted by atoxyl at 1:22 AM on May 2, 2017


More reducing paint to a computer algorithm, no? The placement of paint is either randomized or semi-random and partially guided by human judgement - which puts it in a similar category to some pre-software abstraction.

Or to put it another way, I find a neural network that learns to do Van Gogh pastiche unnerving. A program somebody wrote to produce patterns of a certain kind I don't - it's a concept that goes back at least to analog electronics if not, say, drip paintings. And the guy who made it knows why it works, so one might argue the program is really the art here. But then maybe you just mean you don't really like it, which is of course completely fine.
posted by atoxyl at 1:32 AM on May 2, 2017


Then again, there's always a brush, paints and water. Oh, and a bit of talent...
posted by jim in austin at 2:16 AM on May 2, 2017


As someone who started working in the digital arts software industry not very long ago, I mostly just want to say this is SO COOL. It's a very clever and simple way to get some of the random, chaotic behavior of fluid flow using just simple polygon deformation! We like to use processing in the office for prototyping, because there is so little setup required to create a functional and interactive application around the algorithm you're testing. (For a whole bunch of fascinating and often interactive examples of processing-based art online, check out OpenProcessing.org)

To iterate on what Atoxyl said above, most pro software will be bitmap based (rather than vector) and will indeed be using Computational Fluid Dynamics algorithms to allow you to deposit paint and make it "flow" not unlike real watercolors.

However I could imagine ways of making even this sort of algorithm more interactive, like a traditional painting program, if you could "paint" the base polygon with mouse clicks and then later trigger when you want the polygon deformation algorithm "flow" to start.
posted by Nutri-Matic Drinks Synthesizer at 4:49 AM on May 2, 2017 [1 favorite]


This is cool, thanks for sharing! I've done some very, very basic playing around with generative art in Processing (started by working my way through this book, which I found very clear), so it's fun to see an application that I wouldn't have thought of.
posted by metaBugs at 5:24 AM on May 2, 2017 [1 favorite]


Is this being presented as something that's never been tried before? Surely Fractal Design Painter and programs like it must use some sort of similar algorithmic approach.
posted by lagomorphius at 6:25 AM on May 2, 2017


I got only the impression that it was being done by the author as something they hadn't done before that made for a nice intersection between their watercolor hobby and their programming hobby. Certainly there have been many attempts, more thorough and varied, to model/emulate paint flow in actual art programs, yeah.

More reducing paint to a computer algorithm, no? The placement of paint is either randomized or semi-random and partially guided by human judgement - which puts it in a similar category to some pre-software abstraction.

That's my posture toward it as well, and I think that applies as much to digital painting tools as it does to one-off experiments like this; maybe more so because there's less of a sense of "here's a little experiment I did" with those art packages and instead a more straightforward "here's a toolset, make art with it" mandate.

I've been thinking a lot about tool use in art lately, which is a big ol' rant that'd be out of place here but the core of it as it applies to this stuff is that I am coming to think we tend to very, very artificially separate tools into buckets of different artistic legitimacy based on familiarity and historical cachet rather than any concrete basis.

That e.g. using a computer to model the flow of pigment in water is suspect, while using a paintbrush to deposit pigment in water is considered unimpeachably valid, is a weird quirk of our art culture.
posted by cortex at 7:10 AM on May 2, 2017 [2 favorites]


  Is there a way to plug this into a human controlled graphics program?

I think his deliberately obscure choice of writing this in a Clojure wrapper over Processing pretty much guarantees a “no” answer.

Nice(ish) way of simulating diffusion, but I find his work too cold and algorithmic.
posted by scruss at 7:29 AM on May 2, 2017


« Older A Better Way to Code – Mike Bostock – Medium   |   Music, he tells them, is painted on a canvas of... Newer »


This thread has been archived and is closed to new comments