If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
A Hammer Striking Emptiness 2: Electric Boogaloo
Staring Bodhidharma
Amazing..
Thank you, nice works everyone
Gassho
Washin
sat today
Kaidō (皆道) Every Way
Washin (和信) Harmony Trust
----
I am a novice priest-in-training. Anything that I say must not be considered as teaching
and should be taken with a 'grain of salt'.
Byrne, may I borrow this awesome picture of yours for my e-collection? Thank you
Gassho
Washin
st
Kaidō (皆道) Every Way
Washin (和信) Harmony Trust
----
I am a novice priest-in-training. Anything that I say must not be considered as teaching
and should be taken with a 'grain of salt'.
I have done some work with all of your submissions (except yours Sozan, you hadn't posted yet when I started this -- I will circle back around and work on them soon). I'll be posting them a few at a time over the next day or so. Feel free to just enjoy (or be offended as the case may be) without reading the essay below.
For much of my life I've been interested in emergence, and have been playing with software that runs very simple rules on large datasets to find rules that result in "interesting" or "beautiful" new orders within the data. Generally I find the results most pleasing when working with cellular automata rules in images (but sound is also fun). For anyone familiar with computer science or mathematics, think Conways "Game of Life" or Ed Pegg's "Turmites".
Anyhow, on and off for the past few years I've been focused on processing individual pixels in images using simple rules that adhere to three commandments:
1. The rules must be "local" - for each pixel in the image, the rules can ONLY examine the 8 pixels immediately surrounding it (e.g. to the N, NE, E, SE, S, SW, W, and NW). This means that any interesting higher-order patterns that emerge come from rules that only operate at a local level.
2. Images are treated as "spherical" and "continuous" (e.g. when looking at neighboring pixels at the edge of the image, we wrap-around to the other side of the image).
3. The rules must be "non-destructive". No pixel is ever overwritten, only swapped with a neighbor. This means that no "information" in the image is ever created or destroyed, only re-arranged.
The software I've been working on has a few dozen rules I've found that often produce interesting results. All of the rules compare pixels based on things like hue and brightness, and swap pixels based on the results of the comparison. By way of example, I only ran one rule on Meishin's photo initially to produce this result:
This is "Rule 17" (my numbering as I find interesting ones, no meaning here):
For each pixel in the image, starting at the top left, and looping through:
- If the pixel to the SE of the target pixel is more RED than the pixel to the S of the target pixel, swap the target pixel with the one to the SE.
- Otherwise, if the pixel to the S of the target pixel is more GREEN than the pixel to the SE of the target pixel, swap the target pixel with the one to the SW.
- Otherwise, if the pixel to the SW of the target pixel is more BLUE than the pixel to the S of the target pixel, swap the target pixel with the one to the NW.
- Otherwise, do nothing.
What exactly a given rule does is completely dependant upon the content of the image. Some rules will do nothing on a given image, some will completely rearrange the images in chunks and "streams" of pixels. Like Rule 17 above, which absolutely loved Meishin's photo and produced lovely chunks of color with little streaks through it.
Anyhow, each rule tends to move pixels in a general direction (again, not always, but it tends towards such), so I have them arranged basically like a joystick, and can just sort of surf from rule to rule watching them reorder the target image. Using this technique, I generated 10 images for each of your images, and then picked my two favorites out of the ten.
Gassho,
Sekishi
#sattoday
Last edited by Sekishi; 05-11-2016, 05:49 PM.
Reason: Linked to higher quality version of the image.
Sekishi | 石志 | He/him | Better with a grain of salt, but best ignored entirely.
Comment