Between the Sheets Completed

Here is the video documentation of the project.  Unfortunately, only George and I probably got to see the installation functioning due to several problems.

No Longer Stuck on Stickers. Turned On to LEDs

The stickers don't seem to be working too well for motion tracking.  Their color values are so fickle and highly dependent on lighting.  I did anticipate the lighting to give me trouble, but I'm really questioning the worth of dealing with this trouble.  I think having a light in the curved space will ruin the ambiance of the projection.  I am thinking of buying small LED flashlights and clipping them on a headband (much like Luke's headband, except I plan to buy pre-made LED lights, rather than wiring my own- see the picture.  And, Luke, if you are interested, they make infrared LED flashlights exactly like these.)  I think this will alleviate several concerns with the tracking.  I still plan to use color tracking, and this way I can keep the area dark, I don't have to worry about what colors the viewers are wearing, and the color values I am tracking should stay fairly constant since they are their own source of luminance. 

 

Walmart Nightmare

After missing the first bus, I waited an hour and took the next Downtown, and then got on a bus to Walmart to get the sheets.  It was an hour ride, and it would take another hour to get back.  I knew I needed to make sure I got EVERYTHING I needed.  I wouldn't have another two hours to ride back (plus waiting time).  I bought everything on my list and just as the bus pulled up, I thought, "gee, that's strange, that one sheet looks a bit different shade than the others."  I got on the bus and immediately opened all the sheets up to compare, and yes! they were in fact different.  The thread counts didn't match the color was slightly different and, what's this? The tags actually say a different brand name!  Someone put the wrong sheet in the package!  So now I have two matching sheets and an oddball!  Someone probably bought the sheet, then returned a different one in its packaging.  It seems so wrong.  

I can't use them together because the projection will NOT look the same on them.  But I CAN'T spend another three hours just to go to the Walmart that is only ten minutes away by car.  I don't know what to do but to go around begging anyone with a car to drive me over there so I can exchange the sheet.  I hate having to bother everyone else.  And I wasted my whole day doing this.

Hardware "Goodies"

I just got back from the hardware store with a bunch of "goodies" for my project.  I've got 10ft steel conduit, shelves, a brand new power drill, and much more.  The plan for building the frame is as follows:

The frame/rod will have a diameter just under 15ft.  I will bend 3 segments of the steel conduit using a pressure bender (or rollers) in the sculpture studio.  Then, I will cut small segments of wooden dowel to fit into the end of each of these segments.  I will screw this dowel in place from either side of the conduit.  Once I bring the segments to gallery 120, I will screw the dowel in place inside the end of the connecting segment of conduit.  So, two joints; three segments.  I plan to bend two or three segments for the bottom of the sheets as well.  This will allow me to make sure the "screen" hangs perfectly.  I probably won't joint the bottom segments together since they won't be holding any weight.  Instead, a little duct tape could hold them together.

For the screen:

I am buying three king sized sheets and ripping the stitches out on the sides.  This way, I can use the hem at the top like a pocket and slide it over the rod.  The sheets will have to be sewn/adhered together.  I am waiting for permission to use a sewing machine from the fashion department. 

BIG Problems, simple solution

    While it seemed my equation was working properly, when I increased the size of the matrix so that I could manipulate a longer image appropriate for wrapping around a cylinder, suddenly things were wrong.  The image looked awful and blocky and the image was off center.  I will admit my equation was not quite right, but I had fixed it, realizing that and yet it still was not working.  What could be wrong?!

    After wasting MANNNNNY hours trying all sorts of things and many new equations, I finally figured it out.  I never specified the dimensions for jit.qt.movie and it was defaulting to 320x240.  Using an 840x240 image (a small version for testing purposes), and a matrix with the same dimensions, all the math was coming out wrong, since the equation calls the x dimension as a constant.  BIG problem, simple solution.  I typed in the dimensions for jit.qt.movie and it was all sorted out.  My result is here for viewing. 

Computing Success, Physical Dilemmas

 A conversation with a physics major has brought about a new equation, similar to the others I had tried, but different enough, as the results to the left show.  This seems to be working generally.  I have this patch functioning as a subpatch inside a patch rendering a 3D cylinder with this as a texture.  I am still trying to determine how to delete half of the cylinder so that this matrix only stretches over a semicircular surface, rather than all the way around the cylinder. Once that is completed I will set up the motion tracking to determine how this texture will change with the viewer's position and angle.

    I  want to start constructing the frame for the projection material as soon as possible.  I will need to take some  measurements at the installation site.  In my discussion with a professor from the sculpture department, I was encouraged to consider using rear projection.  The amount of  space and shape of the room will determine if that will be possible.  If frontal projection is the only option, the projection surface will be made with wooden panels instead of cloth, as I was advised is the easiest method.  I was also encouraged to think about how to produce a total immersion, and to consider more projections and the configuration of the installation.  I must somehow balance efficiency and ease with the experiential success of the installation.  I believe my measurements will give me some answers as to the bounds of efficiency and I could then better weigh these concerns.

Spatial Mapping is Not My Friend

I had held out from posting, hoping that at any moment (over the last week), I would have the solution to the problem of the expression needed for the spatial map I am trying to create.  But that moment has not come, though I feel as if I've tried almost everything.  Above is my patch as is.  As you can see, it doesn't work.  The image should look increasingly stretched as you move from the center in either direction, while the center should correspond exactly to the center of the original picture.  I'm quite exhausted at this point and it is getting old spending every free moment I have trying to solve the problem.

Two-part manipulation: Spatial Mapping & 3D

As I have already written two proposals on this artwork, I am simply going to continue to update at this point.  My accomplishment from last week, involving texturing a cylinder will go to use in adapting the video so that it projects undistorted.  However, I find that I cannot use this same method for manipulating the video to account for the viewer's perception of the projection, due to the fact that 3D objects are rendered directly to the video card.  As they are written directly to the video card, there is no matrix output for them, and therefore, I cannot manipulate an image after sending it to the renderer.  So, I must change the video to account for the viewer's perception FIRST, which I have been working on using spatial mapping (as in jitter tutorial 39) and then set that matrix as a texture for the cylinder. 

Maintaining the illusion

At the moment, I am debating whether to require viewers to wear a specific hat in order to track the angle at which their head is turned, which would dictate how the video is manipulated in order to appear straight on to the viewer.  This would provide a more immersive experience, as the illusion would not be broken by a simple turn of the head.  However, it does break the naturalness of the viewer’s interaction with the environment by forcing viewers to wear something they normally wouldn’t.  I have yet to make a decision on which loss is more substantial, and I think, now that my calculations have become more defined, that it will not be a big difference in effort to change the program to account for the angle of the viewer’s head.  Therefore, it is probably best to leave the decision to later, after I have tried the simpler version that doesn’t account for it.  Only by walking through myself do I feel I can make the most informed decision.

 (Yet, I’d be interested in what other people think about which would be more satisfying and would best maintain the illusion, so comments are appreciated.)

Project 1: a short proposal

Altering video for projection on curved surfaces

Having recently done an installation using projectors for my video class, it came to my attention now that I will need to alter the equation I have written down in my notes regarding alteration of video for projection.  The previous equation took into account the way a curved image would appear to a viewer, but did not take into account the way in which light bends.  Unlike the way an image appears on curves, projection size distortion is NOT related to angle specifically, but IS related to proximity directly.  Things become more complicated now and I feel I've barely scratched the surface of the project's application.  I will need to experiment with a projector in order to understand the way the light will bend.

location and structural dilemmas

As I am planning to create a large curved corridor through which viewers will walk, I was presented with the dilemma of creating some sort of curved frame from which to hang the curtains/screen.  But I may have found a solution to the problem that does not include creating a frame at all.  If I can gain permission to use the rotunda of Shafer, I can use the pre-existing curve on the ceiling in order to hang the "screen."  The walls are not perfectly flat, so unfortunately, I cannot hang the "screen" against them.  But the ceiling, interestingly, has a curved ledge.  In terms of my project, I see two functions for this ledge: 1) as a place to install the projectors; and 2) as a place from which to hang the "screen."  Unfortunately, I cannot do this with the same efficiency as a child builds a fort from blankets, simply placing heavy objects on top of the sheet's end to hold it in place.  This would be insecure and dangerous.  So my new dilemma is to find a way to attach the "screen" to that ledge without causing damage to the ceiling (or the viewers!).

Blog 1

As the capstone project should encompass 50% of my grade and workload in the CAR 530 class, I am intending, pending approval, to produce my capstone project as project 1 and half of project 2.  I would complete a smaller project to fulfill the other half of project 2.  This is the best way I can think to satisfy the standard of 50% focus on the capstone project, while not entirely disrupting the format of the CAR 530 course, which is broken up into thirds.

            Assuming that this is the path I will be taking, I feel the most logical thing to do, given the way in which the deadlines for CAR 530 match up with those for the honors program, it to first focus on the parts of the project that involve constructing the environment, including building a frame/screen, video taping (a rough copy of) an outdoor scene, completing a program that manipulates the video properly, and solidifying and implementing a projection plan.  This portion will be project 1.  Adding in the character, setting any parameters for her, perfecting her role, and developing a final copy of the outdoors scene that includes her will count as the first half of project 2.

 

Blog 2

Currently I am looking at panoramas.  Is it possible to make a successful video panorama?  I need the environment of my installation to be seamless.  I thought perhaps I might have more success with a spiral configuration than with the maze configuration I had originally proposed.  However, the idea of a spiral installation presents its own problems.  For example, if I do manage  to make a successfully “seamless” video environment, how can I keep the image in the proper perspective for the viewer?  I want the viewer to feel like he is in the space the video presents, not in a tent of projections.  But I cannot begin to imagine the way imagery distorts in our perception as we move (change points of view) and I am struggling with whom to ask for help on this question.  Someone in the math department, perhaps?  I never could visualize the relationships between the dimensions.  Essentially that is the challenge of this piece, to fabricate the sense of 3D in two dimensions.

            I think I must develop an equation describing the relationship of each point’s position to its…….

 

Well, before I bothered to finish that thought, I scurried away and came up with some notes (not before some failures, of course.  It was all made possible by an index card, haha, on which I was writing on one side to analyze the change in perceived distance between points as a plane is rotated.  My crude drawing was rather unsuccessful, but the lines on the back of the card showed me that perceived distance changes by angle, but NOT PROXIMITY.  Good thing I used one of my lined index cards!  That was all I needed to realize to get started.).  It took me a long time, but I did manage to crank out some equations to get me started on application.  I think they will operate successfully provided that the programs (and the programmer…) are capable of doing what I have planned.  I think the spiral, or perhaps just a large curve, will work much better than the maze formation I had originally visualized, provided I can manage a video panorama.  I think that makes my next step obvious.  Notes are attached.