So you painting it manually? Why not just write/ find voroni noise generator and use it? Even if you can't find it for photoshop, other soft has it (after effects and world machine 2, which is free).
I was playing around with this in GIMP, and if you're geeky enough to care, you can do Voronoi diagrams with two other metrics: the minimum-coordinate metric L(x1, y1, x2, y2) = min(|x2 - x1|, |y2 - y1|), which is the "square" gradient, and the Manhattan metric L(x1, y1, x2, y2) = |x2 - x1| + |y2 - y1| -- to do this one, take the square gradient and use the rotate tool to rotate it by 45 degrees. The example you've used is actually kinda interesting under the Manhattan metric especially. :D
I forgot to add: a proof of correctness is not very hard. Assume your gradient is the exact same layer for all post offices, and that it is an always-decreasing function of the metric you're using. Then the "lighten" modes pick out the maximum lightness, which means that cells share a boundary only when they have the same lightness. Your gradient is monotonic and therefore invertible, which means that the boundary is the same distance from both post offices. That's the Voronoi property. QED. :D
This is an awesome use of photoshop! i didn't knew you could use photoshop that way!
pelegtz 2 months ago
So you painting it manually? Why not just write/ find voroni noise generator and use it? Even if you can't find it for photoshop, other soft has it (after effects and world machine 2, which is free).
izvarzone 3 months ago
@izvarzone Then what would happen to Man's quest for knowledge?
digitalArtform 3 months ago
I was playing around with this in GIMP, and if you're geeky enough to care, you can do Voronoi diagrams with two other metrics: the minimum-coordinate metric L(x1, y1, x2, y2) = min(|x2 - x1|, |y2 - y1|), which is the "square" gradient, and the Manhattan metric L(x1, y1, x2, y2) = |x2 - x1| + |y2 - y1| -- to do this one, take the square gradient and use the rotate tool to rotate it by 45 degrees. The example you've used is actually kinda interesting under the Manhattan metric especially. :D
Drostie 1 year ago
I forgot to add: a proof of correctness is not very hard. Assume your gradient is the exact same layer for all post offices, and that it is an always-decreasing function of the metric you're using. Then the "lighten" modes pick out the maximum lightness, which means that cells share a boundary only when they have the same lightness. Your gradient is monotonic and therefore invertible, which means that the boundary is the same distance from both post offices. That's the Voronoi property. QED. :D
Drostie 1 year ago
@Drostie I'll have to give the Gimp a closer look. Thanks for the info.
digitalArtform 1 year ago
Interesting. Sort-of Worley looking. Thanks.
ronviers 2 years ago
Thanks.
If you invert it it looks pretty similar. It's probably one way of creating Worley noise, if you want to (for some reason) hand place each 'cell'
8760248 2 years ago