Theatre Design & Technology - Summer 1988 - 26

tor. The point at which it crosses the picture plane defines
the vanishing point. The horizon is determined from the
section cutting plane. (In this instance, the projectOr has
been strategically placed on a horizon at the midpoint of
the screen.)
Lines from the vanishing point are drawn to the tOp and
bottom comers of the near edge. The resulting figure is an
accurate representation of the screen as it appears from the
projector. More importantly, however, this shape is also the
corrected projection for this particular projectOr angle and
lens. If a slide of this shape is projected, it will fill the entire
screen, and the edges of the image will be parallel.
What can also be determined from the drawing is the
maximum size possible using a standard 35mm slide
frame. The standard aperture for
a 3Smm slide has a height of
23mm. In the case of Caligula's stage left projection, luck
or good management has arranged that the image size be
3S00rnm x 2300rnm. The scale
is exactly 1: 100. The image size
for the stage right projection is a
little larger, so the scale is a notso-perfect 1: 108. That just
means a bit of time with a
calculator.
Perfectly corrected handdrawn slides can be manufactured from this perspective
drawing (Fig. 3). Obviously,
all the information on a slide
must be in the correct scale
and must be drawn in perspective, but there are no
other major problems. A much
more complete description of
the principles of perspective is
available in Harry Morgan's
book, Perspective Drawing jor the Theatre. 2
The Caligula slides were all3Snml, but there should be
no reason why the same procedure carmot work for other
formats or for the very large slides required by direct beam
projectOrs. The calculations, however, must begin with the
lens data, or with square law infonnation on the size of the
image using a speciRc format at a specific throw di tance.
The Caligula screens were rectangular, but the same procedure can apply to odd shaped projection surfaces.
For the purposes of Caligula, and for trying to verify
these calculations, locating the projectOr vertically at the
midpoint of the screen meant that the distOrtion was in
one plane only. The geometry used here is for two-point
perspective, and is therefore relatively easy to work With.
Placing the projector lower or higher in relation to the
screen would still allow the distortion to be predicted, bLlt
it would necessitate using three-point perspective, which
is JUSt that much more complicated.
This may be very well for hand-drawn slides, bLlt calmot
work directly for photOgraphic material. Objecls thar are
photOgraphed directly onto a slide will not appear in the
proper perspective, and not even the best actOrs will be
able to distOrt their faces accurately enough to compensate. Since the final image on the screen will be two
dimensional, all the material photographed for the projec-

tion must also be in the sanle form. A two-dimensional
scaled graphic of the final image must be prepared. Any
three-dimensional objects that are to appear in the image
must first be reduced to a graphic. That means first photographing the required information on negative film, then
rephotographing the prints onto slide film.
A scaled graphic is still not a corrected graphic, however.
The information must be distorted somehow before it gets
onto the film. In a paper entitled "Photographic Techniques for Creating Projected Effects"3, Merrill Lessley
outlines a method for creating corrected images by photographing a scale model of the set from the exact same angle
and distance [also to scale) as the intended projector. This
theory was successfully applied to the Caligula lides.
Using a copy stand, the graphics
were photographed from the
same relative position as the
point from which they would be
projected. The shape of the
graphic in the viewfinder was
exactly the shape of the screen
as predicted by the perspective
drawing. And all information on
the graphic was viewed in pera
fect perspective.
With a specific distance, hmvever, there is the problem of the
size of the image. For this to
work properly, not only must the
shape of the graphic be correct,
but it must also fill the aperture
of the slide exactly the way the
perspective drawing does. For a
fixed camera lens, and with a
great deal of care and trial, the
required scale could be calculated so that the graphic properly fit the slide aperture. For
Caligula there was an easier solution - a zoom lens. Once the camera position was established, the zoom was adjusted so that the near edge of the
graphic was the full height of the viewfinder. Then both
the size and shape of the graphic exactly matched the
screen in the perspective drawing. The resulting slides
produced in this manner can be projected as correct
images.
In dealing with front projections, that is the end of the
story. BLIt Caligula was done u ing rearscreens. Simply
turning the slide backwards does not work as it would if
the projection angle were perpendicular. If the slide is
reversed, the distortion is compounded rather than corrected. All graphic information must therefore be drawn or
photographed in reverse in order to appear correctly. In
creating the graphics for Caligula, the original negatives
were printed in reverse before being shot ontO slide film.
That's all there was to it.
The process of photographing a photograph allowed manipulation of the graphic for other purposes as well. The
firSt step allowed the objects (in this case mostly facesl to
be photographed from the viewing position of the audience
instead of the projector. Images could then be oriented to
relate to each other, or to peer down OLlt of the screen at the
actor below. Images and actOrs could be seen as pan of a
unified space, thus creating this kind of interaction that

"Since the problem

of distortion can
be viewed as

purely mathematical

one, it must therifore
be possible to solve

it mathematically."

TD&T * SUMMER 1 BBB

Table of Contents for the Digital Edition of Theatre Design & Technology - Summer 1988