Depth of Field Adapter on Steroids

Making these animations helped me understand a few things about Matt's rig (why the diffusion screen is needed, why he used two Fresnel lenses, and why using one is not really an option for this rig), hopefully these will help others.

Here's a GIF of what the Fresnel lenses are doing to keep illumination fairly constant even while focusing. The dotted line is where the focus screen would be (it isn't there). View at original size to see the animation if it doesn't play otherwise.

An ideal 432mm projector lens focusing three points of light at infinity (one in the center, one at each corner/edge), with two identical 432mm Fresnel lenses (also "ideal" lenses in this simulation) condensing light towards the observer camera even as the rig focuses.

Note that when the screen is not present, as is the case for this diagram, the observer camera practically always sees the scene stuck focused to infinity! The observer camera can only capture the rays it can see. A 35mm f/0.4 lens would catch everything, but if you had one you wouldn't need the rig.

Now here's an animation kinda-sorta showing the rig focusing with a diffusion screen present on the same scene.

Diffusion would be more directional than this on an actual screen, but as the previous diagram shows, the Fresnel lenses are directing that diffusion generally towards the observer.

I was also thinking a single Fresnel of an appropriate focal length could be used, but that's problematic:

Oops, this only works the same as two Fresnel lenses when at infinity.  The image formed on the focusing screen gets reduced and illumination at the observer falters so the observer would need to move, zoom, and refocus.