American Cinematographer - July 1983 - 30

3)Wedge

Projection Systems
Today, 3-D movies are being touted
being better than ever. Fact or
hype? Here are a few things one
should know about 3-D projection be¬
fore starting production.
Prior to 1965, virtually all 3-D pre¬
as

sentations

were

with two interlocked

projectors, each running a strip of film
of the two interlocked

from

one

eras

used for

cam¬

filming. Nearly 200 films

have been made with twin-camera

systems (including SPACEHUNTER).
Since twin cameras are bulky,
heavy, and eat twice as much raw
stock, brilliant minds turned to figur¬
ing how to achieve 3-D with one

is

bright horizontal bar through the
picture.
prisms exhibit astigma¬
tism, which may or may not be notice¬
able, depending on design.
Bernier side-stepped the first prob¬
lem by designing three different
prisms, for three theater types. As
such, the prisms didn't have to tilt
a

center of the

much.

The next advance in 3-D
came

projection

from Chris Condon, then of

Figure 2

The

biggest advantage of the splitdesign is simplicity. It takes sec¬
onds to install (replacing the theater's
own lens), and has no knobs or mov¬
ing parts. The image quality can be
lens

camera.

In

1965, 3-D pioneer Robert Bernier

introduced

Spacevision, the first
practical single camera 3-D system.
Out of necessity he also introduced a
corresponding projection system. His
idea has to simply hang two wedge
prisms in front of an otherwise normal
projector. Simplicity has its advan¬
tages-and its disadvantages. The
idea worked but was bulky and re¬
quired an elaborate tubular "jungle
gym'' for support.
The wedge prism principle is quite
simple (Fig. 1). The two 3-D images
(over/under in a single frame) exit the
projection lens with each image going
through one of the prisms, thereby
being refracted (bent) toward the
screen (appropriate polarizing filters
are on the front sides of the prisms).
Adjustments of the distance where
the two images coincide (for a partic¬
ular theater's throw) is made by tilting
the prisms.
The problems with all wedge de¬
signs are:
1) If they are tilted more than 10 de¬
grees or so, keystone
occurs where the. top
the
the

the

anamorphosis
and bottom of
image are different sizes. Since
prisms are mounted in opposition,
two images are distorted in op¬

posite fashion, creating a "bow tie"
effect. This is quite unacceptable as
precise image registration is manda¬
tory for strain-free 3-D.
2) Such prisms, especially when
tilted much, generally produce un¬
even illumination vertically. The result
28

Figure 1

excellent and does not introduce geo¬
metric distortions as do prisms or

Century Precision Optics. In 1968,
having just devised a new single cam¬
era 3-D system called Magnavision,
he looked toward the Bolex (Paillard)
3-D system (circa 1953) (Fig. 2). His
camera put the two images side by
side in a normal frame. So, by putting
literally two lenses in one barrel, he
had the simplest 3-D system.

mirrors.

The main drawback lies in the fact

that,

as

it is

a

lens, different focal

lengths for different theaters must be
stocked. Further, as single film for¬
mats are not standardized (yet!), dif¬
ferent filming systems have different
image spacings. The spacing of the
two lenses must be adjusted for each
film format; an adjustment that can
only be done "at the factory".
In 1971 an independent filmmaker
made
75

an

adult 3-D film and

even

built

projection units for it. According to

the Institute's records, that was the
use of the basic mir¬

first commercial
ror

design (Fig. 4).

Figure 2
The final

design (Fig. 3) was made
by cutting two projection lenses
about in half and putting them to¬
gether. By controlling how close to¬

gether they were (spacing), one
controlled where the

images superim¬
posed. If the center spacing of the 3D images on the film were .425" and

the centers of the two lenses
same,

were

MIRRORS

Figure 4

the

the two images would be out of

perfect registration by only .425". In
practice the spacing is slightly less so
the images actually register at a nom¬
inal distance, keeping the error below
.25" in normal theater configurations.

The two

primary problems with mir¬
systems: 1) Bulk-most will not fit
a Phillips/Norelco or Cinemechanica
(turret) and an occasional U.S. model.
Real problems with mechanical clearror

Continued
AMERICAN

on

page

121

CINEMATOGRAPHER, JULY, 1983
American Cinematographer - July 1983

Table of Contents for the Digital Edition of American Cinematographer - July 1983
