Studio Visit and Interview on Creators Project
Studio Visit Interview on Periscope
Feature on Rhizome
Year of the Glitch is now accepting submissions
Seeking supporters for a chance to glitch Times Square
Retinal Pigment Epithelium…
I was approached by Adam Ferriss, a Los Angeles based artist (check out his tumblr!), about some tips and tricks for circuit bending digital cameras. His work with algorithmic image processing produces images that bear a striking resemblance to those produced by my prepared digital cameras. The photography lab he runs at a college in LA was downsizing their inventory and getting rid of some antiquated FujiFilm FinePix s9000 cameras, and rather than throw them out, Adam decided to hang on to the lot and experiment with circuit bending them.
These things are beasts: fixed zoom point and shoot cameras with the look and feel of a DLSR but without any of the manual controls and flexibility. No wonder they were getting rid of these things!
In exchange for a couple of the less functional cameras, I agreed to help Adam by documenting my deconstruction process. Bonus for you since, now I’m publishing the documentation for public consumption.
Disclaimer: If you’re going to attempt to prepare/modify/circuit bend/disassemble any electronic device, be aware that you are placing yourself at risk of serious injury or death from electric shock; electronic devices may be irreversibly damaged or destroyed (for what it’s worth, it goes without saying that all warranties will be void); if any loss of property or injury occurs, it will be solely your responsibility.
Before opening up the camera, there are a few items we need to have on hand.
- Precision Screwdrivers
- Spare batteries or external power supply (the s9000 uses a 5V supply or 4x AA batteries)
- A bag or containers to place screws and other bits in
- A note book and camera for documenting
- Anti-static wrist band
You’ll also need to do the following to prepare your camera.
- Remove batteries
- Remove the memory card(s)
- Put on and ground the anti static wristband
Now we can begin.
Removing the Screws:
Remove all exterior screws. Like all devices, there are screws in places you wouldn’t think to look. Start with the bottoms, then move to the sides, open all compartments and look for those hidden ones.
Once these are out of the way, you should be able to remove the assembly with the shutter release button and the power and other operation mode switches. Be careful not to pull too hard, like I did, and pull the ribbon connector out of its socket. Fortunately, mine didn’t tear, but you may not be so lucky!
There are still a few screws to be removed before you can open the back panel of the camera. Both of these were revealed by removing the shutter release assembly. One is right next to the strap loop, and the other is just below the flash assembly.
With these two screws out of the way, you should be able to gently coax the back panel off until you encounter some resistance from a couple of pairs of wires. The red and black wires running from the hot shoe attach to a board on the main body via a connector. There’s a speaker on two black wires that attaches to another part of the circuit board via a similar connector. Disconnect these two and the back panel should open like an oven door.
Now that we’ve partially disassembled the camera, and exposed some nice looking innards, we need to figure out if it still works. You can either use the AA batteries or a 5V power supply with a 4.0mm x 1.7mm connector. I used to do a lot of testing for Voltaic Systems and have one of their solar rechargeable V60 batteries around for powering my small electronics projects. Make sure that the main ribbon connectors from the back panel and the shutter release assembly are in place (the speaker and hot shoe wires don’t matter), then power up your camera and turn it on. (hint: check that the battery and memory card doors are closed!)
What’s Inside: Poking About
Now that we have the camera partially disassembled and still working, we can have a look at some of the components inside to see where a good place to start bending would be. Upon first glance, you’ll notice that all the parts are SUPER tiny smd. This is quite a let down, but exactly what you can expect with more contemporary devices. In fact, if you’re opening up cameras released today, you’ll probably find that most all the connections on the integrated circuits are actually underneath the chips and not via pins as with older style ICs!
So what can we mess with? There’s an Analog Devices chip (AD9996) that I can’t seem to locate the datasheet for. There’s something similar to it, the AD9995, which is a 12-bit CCD signal processor. You’ll notice too that there’s a thick connector with lots of contacts. This is the CCD connection (go figure it’s so close to the signal processor).
I actually went a few steps further in deconstructing this camera and found that further disassembling made the system unstable. So, for now, you shouldn’t have to take the camera apart any further to tweek its brains.
How do I mess with it?
Since the pins and connections on this board are so tiny, I am hesitant to solder anything to it. One technique I always go to first is using a saliva moistened finger to poke the sensitive parts and see if anything happens. For capturing, you have two choices: movie or still. You can set the quality settings however you like. If you haven’t already inserted a memory card, now would be a good time.
Other strategies for altering the image is to use a small probe to short circuit adjacent pins on the CCD connector. I found that the right hand side of the connector worked best, and that the series of little smd ICs to the left of the AD9996 gave similar results. When I get in there with a soldering iron to draw out some of those points, I’ll be starting with the ICs and using very fine magnet wire.
So here are a few preliminary images.
I’ll be sharing more of my findings on my year-long glitch-a-day project,Year of the Glitch.
Just uploaded 100+ new images to the DCP Series. This series is still well underway and is branching out into other cameras. Updates to the Olympus PIC series will be ready soon.
Be sure to check out the new additions to the Glitch Textiles project as well. There are currently nine blankets in the collection so far, each featuring a pattern woven directly from an image generated with the prepared cameras of the DCP Series and Year of the Glitch project.
Images and information about three cameras used in recent Year of the Glitch posts and the DCP Series are now public. The Kodak DC280, DC215, and DC200/210, have played a key role in my exploration of hardware and software based image generation/corruption. In the near future, these pages will be updated with more detailed information about specific techniques and circuits involved in creating the variety of images found in both the YOTG project and DCP Series.
Kodak DC 280
Kodak DC 200/210
Kodak DC 215
Going forward, I’ll be looking into patterning some textiles from my own Dither Study series of works inspired by Daniel Temkin’s Dither Studies and featured on Year of the Glitch. If you’re not already following Year of the Glitch, do so today! The project is only 4 followers shy of 1000 and it would be really awesome to make the 1000 follower milestone on Leap Day.
054 of 366
Still remix of For Daniel Temkin (Dither Study 2).
Limited edition signed 20”x30” prints available. Inquire for details.
053 of 366
Stills from a remix of For Daniel Temkin (Dither Study 3).
052 of 366 (full sized GIF click here)
For Daniel Temkin (Dither Study #3)
- Start with a solid color
- Index and dither using an arbitrarily chosen color palette and dithering algorithm
- Zoom 133.33% and crop
- Repeat 2 and 3
051 of 366 (full size click here)
For Daniel Temkin (Dither Study 2)
- Start with a solid color
- Index and dither using an arbitrarily chosen color palette
- Zoom 200% and crop
- Repeat 2 and 3
046 of 366
044 of 366
043 of 366
Images were generated by editing the image data in a hex editor, loading in Firefox, and cropping the screen grabs.
042 of 366
Hex edits performed on the original resulted in a dynamic glitch. In Firefox, the image froze upon reaching the corrupt data. Each time the image was reloaded, Firefox generated a different image.
041 of 366
Hex edits performed on the original resulted in a dynamic glitch. In Firefox, the image froze upon reaching the corrupt data. Each time the image was reloaded, Firefox displayed a different image.
040 of 366
Source: a video consisting of one black frame and one white frame.
Process: iterative re-compression using Quicktime .MOV h264 w/ QT filters and selective manipulation of the raw data using a hex editor.
Concept: successive signal degradation returns to the initial starting point.
039 of 366
A a completely white 4000×4000 px image after seventeen hex edits and 2 format conversions between JPG and GIF along the way. Technique: data bending
038 of 366
A white image 4000px square after 13 edits to the raw data in a hex editor, some selective cropping, and down sampling by 25%. Technique: data bending
(one of a kind 8”x10” signed artist print available for purchase)
Experimenting with making woven blankets out of images from Year of the Glitch. Here are some photos of tests. #32 is featured in there!
There are 4 blankets in this collection. The first four images are two blankets made with a mechanized knitting process. The last two images are two different blankets made using a Jacquard loom.
037 of 366
A white image 4000x4000px after seven successive edits to the data with a hex editor and some selective cropping. Technique: data bending