Zach Dougherty, making reality less stable with animated GIFs
October 15, 2013 2:56 PM Subscribe
Zach Dougherty's hateplow tumblr is a surreal collection of looping animations, of distorted sculptures and glitches in reality.
Is he from Portland? if not he seems to shoot/use a fair amount of places there (newmark theater, Meier and Frank[NOT Macy's!!], portland art museum, etc)
posted by Dr. Twist at 3:17 PM on October 15, 2013 [1 favorite]
posted by Dr. Twist at 3:17 PM on October 15, 2013 [1 favorite]
Dr. Twist, good catch - it appears Dougherty's based in Portland. (Bio at bottom of page.)
The statues turning to look at the viewer are super-creepy.
posted by gingerest at 3:55 PM on October 15, 2013 [1 favorite]
The statues turning to look at the viewer are super-creepy.
posted by gingerest at 3:55 PM on October 15, 2013 [1 favorite]
Awesome.
posted by Rock Steady at 4:22 PM on October 15, 2013
posted by Rock Steady at 4:22 PM on October 15, 2013
Oh man, I saw that statue and feet on Tumblr radar and the GIF didn't animate properly. It's much cooler when it actually moves.
posted by immlass at 4:43 PM on October 15, 2013
posted by immlass at 4:43 PM on October 15, 2013
oh man
the wobbling moon
the postbox that turns around for a quick look
posted by Countess Elena at 4:48 PM on October 15, 2013 [1 favorite]
the wobbling moon
the postbox that turns around for a quick look
posted by Countess Elena at 4:48 PM on October 15, 2013 [1 favorite]
the wobbling moon
the postbox that turns around for a quick look
oh man
-- Basho
posted by uosuaq at 5:02 PM on October 15, 2013 [12 favorites]
the postbox that turns around for a quick look
oh man
-- Basho
posted by uosuaq at 5:02 PM on October 15, 2013 [12 favorites]
glitches in reality? bring 'em on! please!!
posted by oneswellfoop at 5:17 PM on October 15, 2013
posted by oneswellfoop at 5:17 PM on October 15, 2013
These are great.
posted by Sticherbeast at 5:58 PM on October 15, 2013
posted by Sticherbeast at 5:58 PM on October 15, 2013
Honey! That astronaut's in the foyer again!
posted by TheWhiteSkull at 7:31 PM on October 15, 2013 [3 favorites]
posted by TheWhiteSkull at 7:31 PM on October 15, 2013 [3 favorites]
Thanks, that Mary statue thing is going to keep me from sleeping tonight. THANKS.
posted by deathpanels at 8:53 PM on October 15, 2013
posted by deathpanels at 8:53 PM on October 15, 2013
I so badly want to explain, so badly wish to be sidebarred for knocking this out of the park, but may not be qualified depending on what specifically you want explained and how much you and I know about image manipulation and algorithms.
I feel inclined to go out on a limb and take a stab at this, so let's get down to brass tacks and see what we both can get out of this. I think it will be decent and you will feel like you got something out of it. I don't know what you know about algorithms and such so I might over-explain but that's the breaks.
Dude is really good at graphical manipulations and algorithms, and is using additional tools to pre-process some of his approaches beforehand. What do I mean by that? Well some of this basically comes down to writing something like a screen saver or a powerpoint slide transition -- in the first case a standalone computer program that is looping through a set of instructions that progressively make an animated and interesting image, and when it comes to screen savers or programs often referred to in the past as "eye candy," we're not talking about very natural things like a 3D dog ambling across the screen with near-perfect realism. Of course in this case the program generates individual slides that are combined into an animated GIF so we see the result rather than running the program ourselves. Since GIF is being used to handle the animation, the images are limited to 256 simultaneous colors, aka "8 bit" color, and it's a quite a bit easier to manipulate a low quality, animated image like that.
When it comes to classic screen savers we're talking about things like the bouncing magnifying bubble screen saver or a 3D portrayal of flight through space with stars whizzing by. The first example, if you've seen it before, uses whatever you happen to have on your screen at the time as a source / reference image. Then, a computer algorithm decides how to move an imaginary magnifying glass over the image. The more skilled and mathmetically inclined the developer is, the more natural this looks. It might be as simple as "it moves diagonally and bounces across the edges of the screen" which is simple arithematic (take its current x and y position and increase them both, if its position gets too close to the edge then start decreasing them instead) or it might use more advanced math where the magnifying glass appears to bounce pretty naturally with parabolic arcs.
Another algorithm decides how to stretch and disort the magnified area to simulate the effect of a magnifying glass.
The first two examples are like this. The background is probably taken from a scene with no statue at all and then the statue is superimposed. But maybe not -- it's just easier that way, as you don't really need any great photoshop skills to accomplish that vs. the alternative, where the statue really was in the scene and the photographer had to seamlessly capture the background behind the statue. The bottom line is that we have an exact image of the statue itself to work with, and an exact image of the background, so we have enough information to superimpose the statue on the background and then poke holes that show us the background behind it. Not only that, but it's not really a simply picture of a statue, it's been turned into a 3D model, or it started out as a pure 3D model. It's a bunch of really tiny polygons that when assembled create the image of a statue but because they are tiny polygons being arranged a specific way to look 3D on a 2D plane, we can execute an algorithm similar to a powerpoint slide transition that no only erases the face in a circular, outward pattern, but lets us see depth within the statue.
The third image basically looks like a 3D model of a sculpture that was then broken apart using 3D illustration and animation tools. It seems less "algorithmic" other than the algorithm rotating it side to side and is more likely the product of tons of manual manipulation in a 3D studio.
The fourth and fifth are my favorites. We start by taking a real sculpture and creating a 3D model of it using very tiny polygons. With the digital, sterile information describing all of these tiny polygons we can actually "average them out" into much larger polygons, creating a very distorted "3D-pixelated" image that can quickly become unrecognizable. So an algorithm is cycling between "show a lot of polygons and detail" and "show a very blocky weird-ass image." With some intentional jittery-behavior where it jumps to "a lot of polygons" to "very few" without easing between extremes, which causes it to look disjointed and glitchy in certain frames rather than 100% smooth and transitional.
This same concept allows highly detailed computer games to adjust their performance on slower systems by drawing fewer, but larger polygons to represent a 3D object.
The final one is a little of everything combined -- superimposing a 3D model over an image, using a model of a real statue but based on big polygons to create an uncanny valley effect, some rotation, and then possibly some "ray casting" with imaginary rays of light being rendered hitting the nose and symmetry line down the face, though there are other explanations for that I'm sure.
Ta Da!
posted by lordaych at 1:12 AM on October 16, 2013 [5 favorites]
I feel inclined to go out on a limb and take a stab at this, so let's get down to brass tacks and see what we both can get out of this. I think it will be decent and you will feel like you got something out of it. I don't know what you know about algorithms and such so I might over-explain but that's the breaks.
Dude is really good at graphical manipulations and algorithms, and is using additional tools to pre-process some of his approaches beforehand. What do I mean by that? Well some of this basically comes down to writing something like a screen saver or a powerpoint slide transition -- in the first case a standalone computer program that is looping through a set of instructions that progressively make an animated and interesting image, and when it comes to screen savers or programs often referred to in the past as "eye candy," we're not talking about very natural things like a 3D dog ambling across the screen with near-perfect realism. Of course in this case the program generates individual slides that are combined into an animated GIF so we see the result rather than running the program ourselves. Since GIF is being used to handle the animation, the images are limited to 256 simultaneous colors, aka "8 bit" color, and it's a quite a bit easier to manipulate a low quality, animated image like that.
When it comes to classic screen savers we're talking about things like the bouncing magnifying bubble screen saver or a 3D portrayal of flight through space with stars whizzing by. The first example, if you've seen it before, uses whatever you happen to have on your screen at the time as a source / reference image. Then, a computer algorithm decides how to move an imaginary magnifying glass over the image. The more skilled and mathmetically inclined the developer is, the more natural this looks. It might be as simple as "it moves diagonally and bounces across the edges of the screen" which is simple arithematic (take its current x and y position and increase them both, if its position gets too close to the edge then start decreasing them instead) or it might use more advanced math where the magnifying glass appears to bounce pretty naturally with parabolic arcs.
Another algorithm decides how to stretch and disort the magnified area to simulate the effect of a magnifying glass.
The first two examples are like this. The background is probably taken from a scene with no statue at all and then the statue is superimposed. But maybe not -- it's just easier that way, as you don't really need any great photoshop skills to accomplish that vs. the alternative, where the statue really was in the scene and the photographer had to seamlessly capture the background behind the statue. The bottom line is that we have an exact image of the statue itself to work with, and an exact image of the background, so we have enough information to superimpose the statue on the background and then poke holes that show us the background behind it. Not only that, but it's not really a simply picture of a statue, it's been turned into a 3D model, or it started out as a pure 3D model. It's a bunch of really tiny polygons that when assembled create the image of a statue but because they are tiny polygons being arranged a specific way to look 3D on a 2D plane, we can execute an algorithm similar to a powerpoint slide transition that no only erases the face in a circular, outward pattern, but lets us see depth within the statue.
The third image basically looks like a 3D model of a sculpture that was then broken apart using 3D illustration and animation tools. It seems less "algorithmic" other than the algorithm rotating it side to side and is more likely the product of tons of manual manipulation in a 3D studio.
The fourth and fifth are my favorites. We start by taking a real sculpture and creating a 3D model of it using very tiny polygons. With the digital, sterile information describing all of these tiny polygons we can actually "average them out" into much larger polygons, creating a very distorted "3D-pixelated" image that can quickly become unrecognizable. So an algorithm is cycling between "show a lot of polygons and detail" and "show a very blocky weird-ass image." With some intentional jittery-behavior where it jumps to "a lot of polygons" to "very few" without easing between extremes, which causes it to look disjointed and glitchy in certain frames rather than 100% smooth and transitional.
This same concept allows highly detailed computer games to adjust their performance on slower systems by drawing fewer, but larger polygons to represent a 3D object.
The final one is a little of everything combined -- superimposing a 3D model over an image, using a model of a real statue but based on big polygons to create an uncanny valley effect, some rotation, and then possibly some "ray casting" with imaginary rays of light being rendered hitting the nose and symmetry line down the face, though there are other explanations for that I'm sure.
Ta Da!
posted by lordaych at 1:12 AM on October 16, 2013 [5 favorites]
All of my qualifications and warnings had more to do with my own limited understanding -- I grok quite a bit of this and spent many pre-teen nights away writing little screen savers that do cool things typically with no reference image (or would gradually destroy the original image entirely), just sequences of iterating processes deciding what colors to place on the screen...but never evolved myself into really using higher math and 3D.
posted by lordaych at 1:21 AM on October 16, 2013
posted by lordaych at 1:21 AM on October 16, 2013
The jupiter one is a nightmare of mine. I'm walking down an unfamiliar street in an unfamiliar city. I turn the corner and...there it is. Spinning and staring at me.
posted by Our Ship Of The Imagination! at 10:17 AM on October 16, 2013
posted by Our Ship Of The Imagination! at 10:17 AM on October 16, 2013
hows it done ? - using blender, maya, 3dsmax etc edit, not forgetting after effects, video copilot is your friend.
posted by sgt.serenity at 12:25 PM on October 16, 2013
posted by sgt.serenity at 12:25 PM on October 16, 2013
As I look through the progression of his tumblr, I get that he's a photographer first and computer modeler second.
This might help show how the "breathing" shots work.
posted by RobotHero at 6:27 PM on October 16, 2013
This might help show how the "breathing" shots work.
posted by RobotHero at 6:27 PM on October 16, 2013
« Older Cockroach husbandry in China | Tyrant, mass-murderer ... editor. Newer »
This thread has been archived and is closed to new comments
posted by craven_morhead at 3:09 PM on October 15, 2013