How to populate shadows with content

If I hold my hand up, in front of a light, it casts a shadow on the wall. My hand (my image of it) has lots of details. The shadow of my hand (my image of it), is empty of such details.

I had this idea in the late eighties (while at art school) of how to populate a real shadow with an image, while outside the region of the shadow would be another image. I never implemented it due to lack of resources. But here is how I was going to do it.

It involved substituting the wall with a translucent screen. On either side of the screen would be a video projector projecting an image. A projector on one side and a projector on the other. When you held up your hand  in front of one of the projectors, it would cast a shadow. Inside the shadow would be an image and outside the shadow would be the other image.

A computer would operate on a source image (a video signal) splitting that signal into two signals, but in a completely random way.  One signal would go to one projector while the other signal would go to the other projector. Now each signal on their own was just noise (no signal) but when recombined on the screen the result would be the original signal. This was inspired by Neils Bohr's deconstruction of Einstein's EPR experiment and Plato's parable. Anyway, if you were to put your hand in front of one of the projectors, the result would be a shadow, but in the domain of which, would be noise. Everywhere else would be the signal.

Now if the projection occurs at a fast enough rate (eg. 50 herz) there is a perceivable integration which can occur in the interval between one frame and the next (or previous). So two otherwise random patterns, when integrated, would tend to cancel each other out - which is why noise looks sort of empty. But if the circuit was programmed to produce successive psuedo-noise patterns, that when integrated, formed an image instead, then in the domain of the shadow an image would be produced. Outside the domain of the shadow the alternative image would be reconsituted. The outside of the shadow would occur as a function of spatial reintegration on image A, while the inside of the shadow would occur as a function of temporal integration on image B. There would be no bleeding of one signal in the domain of the other.

The arm placed in front of the projector would act as a physical switch between these two images.


Now the alternative way of achieving a similar effect, but doesn't tap into Bohr's deconstruction of EPR is the following:

To projector 1 is sent the image: A-B
To projector 2 is sent the image: B.

Outside a shadow the images will combine to produce: A = A-B+B
Inside a shadow (with a hand in front of projector1) you have only projector2:  B

However, on the surface of the hand will appear image A-B.

In the random version, on the surface of the hand won't appear an image (or signal). What will appear instead is noise.

Carl