The Imagined Materializing
(Notice how I changed the name from "Feedhorn Camera" to "Light Pipe Array")
Detail of the film-plane end of the array. Notice that both the plastic mesh and the front cardboard baffle are warped into a semi-spherical shape by the fact that the grid of holes were all drilled perpendicular to the surface of the material. I drilled out each square in the plastic mesh to permit the plastic straws easier penetration through the grid openings. It took a lot of adjustment, using thin brass guide rods and tubes, to get the rear surface of the array anywhere close to being planar. As indicated in my design notes, if the film plane is right up against the end of the straws, each "pixel" should be a sharply focused circle, whereas with some spacing between array and film they will blend into more of a continuous image.
Here I placed the array facing out my kitchen onto the backyard, and captured a photo with my digital camera. You can begin to see some semblance of detail, like the green grass and blue sky. But the silver gelatin image will look different, other than being in monochrome, as I expect most of the light traveling down each tube will be averaged at the film plane into some spot of intermediate tone.
When looking through the Light Pipe Array the thing that strikes me is the angle of view. Designed to be 50 degrees of horizontal coverage, when peering into the edges and corners straight-on, they appear to be just as brightly illuminated as the center. It remains to be seen how well this holds up when put to the photographic test, however. In this image, I've taken a digital snap looking into the upper right corner of the array.
Finally, this image shows, out-of-focus and over-exposed on the left, a portion of grass and cement in front of it, while you can make out, in the array to the right, the same sense of green grass on top and cement underneath. Note that the Light Pipe Array, unlike other camera lenses and pinholes, does not invert the image at the film plane, but that it remains oriented top up.
Post-Script: As a final note, I didn't make this prototype to the degree of exacting specifications implied in my sketch journal notes. So I expect lots of image field overlap between neighboring pipes, and consequent reduce contrast. All of this is a bit moot with a pixel resolution of only 22 by 22.
A larger version might be designed to fit an 8" x 10" format, which would yield an image almost as "sharp" as a 100 by 100 pixel avatar image. Finding tubes smaller in size than these straws (about 7.5 to the inch in diameter), at some economic price, remains a challenge to upgrading this prototype into a camera with some decent performance. I'm sure the Military Industrial Complex could cough up some, however, but I don't have the war chest to reciprocate. So coffee stir sticks will have to suffice for now.
I can also imagine this idea could be shrunk down and fit onto a digital image sensor, to replace a conventional lens, with each light pipe the same diameter as its corresponding pixel. Perhaps we'll see something like this in the future; if so, you heard it here first.
UPDATE 8.29.2013: Here's a scan of one of the first resulting images from this prototype Light Pipe Array camera, exposed onto Freestyle Photo's Arista-brand grade 2 RC photo paper and inverted in Photoshop. It helps to squint while viewing the image, or step back a bit from your computer monitor.
The exposure was made by placing the LPA into a large plywood box camera, facing backwards towards the rear lid, and using the lid as a crude shutter. Self-portrait under indirect north-facing daylight, several seconds exposure time.
I initially calculated the exposure using the tubes' length and diameter to approximate a pinhole aperture, but the resulting exposures were way over-exposed, most likely because, unlike with a pinhole or glass lens, the light doesn't diverge at an inverse-square-law rate, but reflects down the inside of each light pipe. I ended up applying an exposure adjustment to my light meter reading whereby I rated this prototype LPA at about f/11 aperture.
You can make out the white smock I was wearing, but finer details are not possible, due to the diverging angle-of-view of the array resulting in reduced resolution at the center.
Further indoor testing under indirect daylight showed some further exposure attentuation problems (under-exposing the images), possibly caused by either absorption and/or scattering of the light inside each tube, or the house's surfaces warming the daylight, shifting the spectrum toward the red where the paper negatives are less sensitive. Careful calibration tests are therefore required to determine optimal exposure under various light conditions.
My second prototype camera will be of 8"x10" format, using a parallel-pipe array where all the light pipes point forward in the same direction, optimized for close-up facial portraits. More details to follow in a future blog post.