But as I mentioned we had these thin slices of books, one of which was the edge of this shuttle launch.  Long story short I noticed that when this slice was fanned out and laid flat, an image of the coastline appeared almost in proper perspective, as if you were looking across the length of the beaches.  Kind of a neat thing.  Anyway, below is a comparison shot between the flipbook SLICE and one of the images from the book.  You can clearly see the shadow from the booster plume.

THESIS PROJECT PROPOSAL* updated

Project Title: to be determined

Project Summary:

An interactive fiber optic installation: Movement of viewers through the space will be mapped by computer vision techniques and projected through a three-dimensional array of fiber optics. The intended effect is that “stars” will create a dynamic aura around any individual within the space.

Project Description:

A black box space (quadrant of the Raw Space) will be sealed and capped with a 3’ grid of aircraft cable at a height of 11’. Black fabric will be suspended from the cable and fiber optics of various lengths will be threaded through to create a 3-dimensional array of points of light, representing the night sky. Light to 3 bundles of fiber optics will be delivered via 3 high-lumen projectors. Each projector will be fed a video feed taken from one of three positioned surveillance cameras with infrared filters and illuminators. Those cameras will be run through Max/MSP and will process the images so that any body moving within view will be translated to a white blob against a black background. As an object(s) is tracked, the white image will feed groupings of fiber optics specific to the object’s location. Each projector will map a different axis of view so that the object is accurately represented three-dimensionally with the fiber optics. Essentially, as a person moves through the space, their mass will activate swarms of “stars” surrounding them. Static constellations will also mark locations in the room so that it does not become completely dark.

Several timer settings will be built in to determine the length of decay of the lights; the more people there are, the more rapidly the lights will fade and vice-versa. These timers will be implemented according to time of day and how active the space is in terms of viewers. There will also be playback of captured footage set on a timer, so that the “star fields” will morph “on their own” if no viewer has activated the space for some time. There will also be trigger-able “hot-spots” of “star formation”, rather, locations that trigger sequences of light patterns.

Methodology:

Much of the initial groundwork in terms of planning has been developed in a previous version of this installation. (Space and Place, Fall 2009) The technical issues, such as programming in Max/MSP and space design specifications are constantly in positive process, leading towards a smooth implementation during install week.

Deliverables:

I will be presenting technical progress in staggered bi-weekly meetings with both of my Thesis advisors, Paul Catanese and Jim Ferolo. At each point I will have another stage of programming and prototyping complete so that the installation of this project will have as little “design-on-the-fly” as possible.

PROTOTYPE ONE:

What you see here (if you don’t listen to the audio) is a 4′ cube model of my thesis installation. The actual space I have to work in is 16′ x 16′ x 12′ (with a 10′ ceiling once complete). One of the best ways to figure out if something will work is build a smaller prototype. If you read the proposal, then you’ll have an idea of what this is. But to reiterate: I am using computer vision within the space to track people’s movements and then translate that visual data to fiber optics. So when people move through the space, swarms of “stars” follow them and surround them. The running metaphor for this piece is “the viewer as dark matter”.

So the technique essentially uses Z-axis and XY-axis cameras to track movement. The XY will track what the position the viewer is at while Z will determine “how big” to make the “blob”. This is done by turning the viewer’s form into a white blob against a black background. That video feed from the XYcam is then sent through 3 projectors, each feeding one of three layers of fiber optics in the space. The plan for Zcam is to just use tracking data to decide how big to make the blog on each layer. I’ll need to test that out this next week…

Anyway, this prototype was mostly for testing density of the fiber optics and sensitivity of the camera and light movements. At first I was using my laptop’s camera just to make sure that the movement translated to light well enough, which as you can see it did. I played with the delay a little bit to make it more immediate and legible for testing sake. Once in the actual space it won’t be so seizure inducing. The other thing I tested with this prototype was the XYcam, mounted on the top. The IR emitter and filter weren’t setup so the movements just weren’t tracked as well. That is another item on the task list for this week.

The last issue that I wanted to explore was the lightbox thing. I don’t believe I will need to make any fancy angled mirrored lightbox. In fact, doing so would throw any sort of fiber mapping into disarray. Any mirrors between the projector lens and the fibers will throw the light all around, making it impossible to maintain any sort of cohesive coordinate system. Fuck. SO, I figured out how big I need to make each bundle and how far away from the lens it could go. Lightboxes will still be made, just not with any mirrors. Another item for this week’s task list.

It is also entirely possible that a 2 camera system will be overkill. Well, not overkill but unnecessary. Once I get three projectors doing a pass through I can test each bundle layer to see if I can map them without any Zcam data. Either way, I’m going to try my damnedest to get 2.0 up and running. And then the sound component still needs to be worked out….once again, fuck.

PROTOTYPE TWO:

Oh, where to begin… How about where I left off: Prototype 1.0, where I was mainly testing for fiber density and tracking reactivity. None of the fibers were mapped and only a small portion were threaded. In Prototype 2.o (seen above), I began by building a “light box”, or rather an apparatus with which the fiber bundles would be attached to the projector. Since the image coming off of the projector is rectangular, the ends of the fibers had to be arranged similarly to ensure more accurate mapping. Originally, three projectors (and three apparatuses) were the plan until I, um, sort of ran out of fiber…not completely…but filling up one apparatus used just over 2/3rds of the fiber I had on hand. So at this point I will definitely need more…

Moving on. Once I had all the fibers in place on the apparatus/projector monster, I began mapping fibers to specific quadrants of the prototype’s physical space. In order to discern specific fibers I reset the background subtraction in the Max patch and placed a large object in the location to be threaded. That way the fibers needed for that location were illuminated. It takes a good amount of time to thread the fibers so about half of each bundle was not mapped…otherwise the small space would be WAY too dense with fibers to get any sort of accurate testing. As it is, it seems that too many were added because the camera can detect the swaying of the fibers after my hand left them. That is an issue that can be curbed when the physical spacing of the fibers grows and they essentially become “smaller” in the camera’s eye; all of this will happen once the final installation commences. Until then a simple lowering of the output resolution (big white pixels) does the trick. (please excuse the tense-shifting…I’m in the middle of the project so a mix will have to do…)

Overall the mapping works, even if in a rudimentary way with this prototype. The camera basically detects where movement is and translates that image to the fibers, which are placed where the movement happens…uh. It almost sounds like I’m talking in circles…Anyway, it works. There are some buggy issues, once of which mentioned above, that need to be worked out still. Things like negotiating vertical mapping, fiber density/distribution in the space, making a better apparatus for the fiber bundles, tweaking the patch so it’s user-friendly…oh. Duh. And sound. I am planning on doing sound with this piece as well, similar to that of a couple earlier projects. The starting thought is that there will be an ambient track that shifts very subtly in response to general movement and a “twinkle” track that responds to very specific movements. THAT is what I really have to work on this week.

I should also mention, if I haven’t already, that both of my advisors, Paul Catanese and Jim Ferolo have been highly supportive and tremendous in helping to navigate a number of issues that arise during this whole process. Originally I had planned on using 2 cameras and 3 projectors; one overhead and one side camera with the overhead multiplied into the 3 projectors. Since I haven’t been quite able to figure out how to do the mapping with side camera data, I’ve chosen (with guidance) to use the overhead alone and physically map the fibers vertically. The effect will be pretty close to the intention…I hope…

Also, since I don’t have enough fiber for 3 projector apparatuses, I will only be using 2 projectors. There is still going to be more fiber optic hanging in the space than there was in Waking the Invisible. Even more so because when considering the actual installation space, the fibers don’t really need to fill it completely. In actuality, they will only need to fill a 6′ tall section (beginning at 2′ from the floor) that is 10-12′ square, basically 600-800 cubic feet as opposed to the 2,000 plus to fill the whole room. Yikes. What’s also nice about that realization is that the installation method changes. Jim and I met in the space the other day and were discussing ways to make the process less of a pain in the ass and a time-sucker. Instead of aircraft cable stretched, fabric laid over it and then me on a ladder for a week straight threaded the goddamn things, I’ll make a 12′ sq PVC grid of 4′ sections, stretch the fabric across and thread the fibers while the thing is still at floor level (or just above) then use a pulley system to hoist it to height where it will then be secured! Fucking brilliant! It’s lighter than 2x4s, it wont droop as much as cable because of the center pulley and side anchors and it might only take a couple of days to thread and hang!! Woot! On to it then!

MAX ONE:

More MaxMSP testing. This time around I was just testing the IR emitter and filter setup. Now it can see in the dark! In the next few days I’ll have some footage of this type of video feed generating light patterns in the fiber optics, unmapped and hopefully mapped (reasonably enough)…

MAX TWO:

MAX THREE:

More progress on my Max patch. Still a little bit more to do, like multiply the triggers and get the network setup…

MAX FOUR: