What happens when you take a photo of a photo over and over again, recursively?
How is this filmed?
I used my iPhone 6s Plus to take the pictures, and my MacBook Pro 15 inch with Retina Display for displaying the rendered images. I mounted the iPhone 6s Plus with the Zilu Universal Car Phone Mount on a rotating small mirror, so that I could adjust the positioning of the camera precisely.
After taking a picture from my iPhone, IFTTT (link to the specific applet) automatically backed up the photos to a specific Google Drive folder, which I then opened on my computer to take the next picture.
Why does the image mutate?
When each photo is taken, there are small perturbations that slightly mutate the image:
- Colors. The colors eventually trend toward RGB values, caused by the RGB pixels in the computer screen.
- Human Error: Despite popular belief, I’m not perfect. It was a huge challenge to get the camera aligned to the image on the computer, and there were always parts cut off, or a slight angle which eventually compounded and warped the image, even after tens of minutes of adjustment.
- Brightness: iPhone cameras try to anchor to “middle brightness” to determine how dark or light to make the image. You’re probably used to tapping the iPhone screen and seeing a square appear, which is how the camera calibrates brightness. In this case, the brightness calibration is generally either making the dark parts of the picture darker or the brighter parts brighter.
- Stripes/dotting: There’s a bit of noise from the computer screen that appears in the photos, in the form of pixelation or stripes.
- Glare: Light reflects off the monitor and introduces a bit of a fuzzy effect each time a new picture is taken.
- Miscellaneous: Could be anything else — dust on the screen, slight jitters as a I press the iPhone, perhaps a ray of sunshine happens to appear through the window just as I press the camera button.
So, what do the results mean?
This experiment is obviously not an accurate representation of exactly how the brain deals with memories, but an interesting thought experiment and investigation into the mutating forces of recursion and entropy.
It’s beautiful and fascinating to see the RGB colors of the computer screen come to life in the imagery, and the transformation of the image to the blobs is mesmerizing. For me, it’s a reminder that, despite the “destructive” nature of entropy, there’s such beauty in chaos.
The magnification of even the slightest perturbations was remarkable. In this ice cream photoshoot, my camera was very slightly tilted to the right, so that the left side of the image would get cut off. This effect compounded and eventually caused the whole subject of the photo to shift to the left. In memories, a similar effect occurs — a subtle negative thought or suspicion can compound over and over again, consuming the memory itself.
What other natural processes could this recursive photography symbolize?
- When oral history is told from one story-teller to the next, how does the history change?
- When cultures reinterpret and religious texts from one generation to the next, do they end up reflecting more and more of the culture itself?
- What happens when you translate a story from English to Spanish, then back to English, then back to Spanish?
- Perhaps even the force of evolution can be considered a sort of recursion.
Other Objects: My Face
I was curious to check out this effect on other types of subjects. I noticed that the colors still strongly trend toward RGB values. Quite fascinating to see my own face morph into a beautiful array of colors.
Other Objects: Hand Soap
In this example, I really wanted to investigate the final equilibrium of the photo — does it ever reach a point where it stops changing? This photoshoot took the longest, but I eventually got to a point where I viewed a plain blue screen, which, after several photos, stopped changing.
What is the “final equilibrium” of our memories?
If you enjoyed this article, please show your appreciation by clapping 👏 below, and check out some of my other similar projects: