TEACHING SHOT SELECTION AND
COMPOSITION THROUGH COMPUTER SIMULATIONS
1998 UFVA Conference
Michael F.
Korpi
Description of Project
Purpose
The purpose of this project was to develop a series
of computer simulations to improve the scope and quality of
instruction in CST 2385 Production Methods by strengthening the
portions of the class dealing with visual composition and
shot-by-shot breakdown of a dramatic scene.
Objectives
In order to accomplish the above purpose, there were
two objectives:
- to develop a set of three-dimensional computer
environments (virtual worlds) in which students can maneuver a
virtual camera to obtain shots appropriate for telling an assigned
story,
- to modify the course (CST 2385) to take advantage
of these virtual worlds (both for in-class presentation by the
instructor and exercises which students can pursue on their own
time in the computer labs).
Benefits
Two key aspects of television and film are (1) the
frame, which defines and limits what can be seen, and (2) the
sequence of related shots which is typically used to present an
event and/or a location. The "Hollywood style" strives to make these
aspects of film and television manipulation "invisible" to the
audience. Because of this invisible nature of the technique, the
basic manipulations can be difficult to teach, and very difficult
for some students to learn.
The typical approach is to explain the technique in
a lecture format, present some examples in class, and then send
students out with cameras to complete a series of exercises. The
main problems are that (1) time and equipment limitations mean that
students are able to do only a few exercises, and (2) there is a
large time gap for most students between both the instructor's
explanation of the assignment and use of the camera, and between use
of the camera and feedback from the instructor.
A computer simulation wherein a virtual camera can
be maneuvered about will greatly improve the instructor's initial
presentation of the concept (this portion of the course could even
be taught in a computer lab), but more importantly, it will improve
the efficacy of the student exercises. First of all, more students
will be able to work at the same time because there are more
computers available than there are cameras. Secondly, more complex
exercises can be assigned because students will not be limited to
actual physical locations. Finally, and most importantly,
explanations, guidance and feedback can be much more effective
because temporally they occur in close proximity to performance of
the exercise.
Procedure
The first step was to identify previsualization
software that is capable of creating a simplified but convincing
world (especially the people inhabiting this world), capable of
adjusting the shape of the frame and the angle of view (that is, how
much the camera can see horizontally), able to print any view,
capable of interacting with other software, affordable, and able to
run on the systems in Baylor's computer laboratories (Macintoshes
and Windows PC's).
The software selected is called Virtus Walkthrough
Pro and it is available from Virtus
Corporation, Cary, North Carolina. It meets all of the above
requirements, and it has the additional benefit of being able to
export its virtual worlds in the Virtual Reality Modeling Language
(VRML) format which is used in web browsers. There are many other 3D
design or animation packages which met most but not all of the
requirements. The ability to change angle of view turned out to be
the single most important distinguishing characteristic between
Virtus and alternative software packages.
Figure 1 Virtus Walkthrough Pro screen
The software presents the user with a focal length
(or angle of view) control on the left. In Figure 1 above the focal
length is set to 15mm. Then there are several "Views" available:
Top, Bottom, Right, Left, Front, Back, and Walk. Top and The Walk
View is the most important because it is the one in which students
will navigate through a three-dimensional environment and perform
the assignment. The other views are all design views used mostly to
manipulate the objects in the environment.
The Walk View's aspect ratio (ratio of width to
height) is selectable. In the above illustration the aspect ratio is
1.33:1. The following images illustrate the same scene as viewed in
2.35:1 (Figure 2), 1.85:1 (Figure 3), and 1.33:1 (Figure 4). It is
quite obvious that the aspect ratio can make a big difference in how
the scene can be presented.
Figure 2: 2.35:1 -- Wide-screen Movies
Figure 3:1.85:1 -- Most Hollywood Movies
Figure 4: 1.33:1 -- Television
The second step was to create scenes and stories
appropriate for teaching composition and shot selection. The scenes
begin at a very simple level and gradually get more complex. There
are several scenes which will be used to demonstrate Virtus
Walkthrough Pro and how to navigate in its worlds; then students
will begin working on their own. The student assignments make use of
the following worlds:
One Person
Most of the familiar shot types (e.g., LS,
MS, CU, etc.) are based on the human figure. It makes sense for
students to begin by composing shots of one person. This one person
could be located in limbo, but this assignment places the woman on a
large level surface which extends far enough to form a natural
horizon. The horizon is important because it tends to lead the eye,
and therefore it can be used to strengthen a shot.
On an unobstructed level surface shot from a level
camera, the horizon line will cut through the person or object being
photographed at the same height as the camera. For example, in
Figure 5 the camera is at waist level; in Figure 6 it is at eye
level; and in each case the horizon line bisects the woman at camera
height. This holds for any camera-subject distance (e.g.,
Figure 7).
Figure 5 Camera at waist level
Figure 6 Camera at eye level
Figure 7 MCU Camera at eye level
Two People
This world introduces students to the more complex
task of dealing with two people. Common shots like over-the-shoulder
(OVS), the two-shot (2S), and the plan Americain or
"Hollywood shot" have developed to deal with the interaction of two
people. What matters most is the relationship of the characters
within the frame. In narrow aspect ratios like 1.33:1, the
characters must be placed closer together. In wider aspect ratios
like 1:85:1, there can be more space visible between two people
while still maintaining a pleasing composition. This can be a help
or a hindrance depending on the demands of the dramatic
situation.
Figure 8 OVS in 1.33:1 aspect ratio
Figure 9 OVS in 1.85:1 aspect ratio
Objects
Some shots do not include people. There is a more
limited range of shot types for objects than for people. Generally
extreme long shot (ELS), wide shot (WS), and close-up (CU) serve for
the description of shots of objects.
The "Objects" world is intended to give students
practice in composing shots of a variety of things which vary
considerably in size, complexity, and context. Obviously, "WS
telephone" is different from "WS BMW" or "WS beach cabin." In this
world students begin to get the idea that the surroundings matter as
well.
 
Figure 10
Figure 11
A Baby in the Road
In this world students get to work with some of the
most powerful elements in motion pictures: conflict, parallel
editing, and the camera's power to manipulate depth. The situation
borders on the outrageous or absurd, but the inherent tension keeps
students' attention and helps them think about how the same is true
for the audience. A typical construction of this scene would cut
back and forth among several simultaneous actions: the baby
crawling, the mother not noticing, the oncoming truck, and the
jogger.
 
Figure 12
The camera can manipulate the perception of depth in
order to increase the threat of the truck each time the baby is
shown. This is done by decreasing the angle of view in each shot
while simultaneously moving the camera farther away in order to keep
the baby the same size in the frame. The result is to magnify the
background object, in this case the truck.
Figure 13
Even though the baby and the truck are exactly the
same physical distance apart in each shot, the truck appears to be
dramatically closer each time. When the students make this
manipulation themselves, they know the truck has not moved,
but they can clearly see the impact of this manipulation. They need
no more convincing that the camera can lie.
Shot Selection or "Coverage" of a
Scene
In shooting feature films or dramatic television one
of the most important functions of the director is to deliver
appropriate "coverage" of the scene to the editor. This coverage
consists of shots from several to many different camera locations
covering overlapping and repeated action. Besides the shots
themselves, this coverage of the scene is reported to the editor by
means of the "lined script." For each printed take a vertical line
is drawn on the script to indicate the duration of the shot. This
line is straight as it passes through dialogue for characters
visible in the shot -- that is, on-camera dialogue. The line is
"squiggly" as it passes through dialogue delivered off-camera. Each
line is labeled with the scene/setup/take information (e.g.,
45C-3 is scene 45, camera setup C, take 3). A sample lined script
page from the television series Highlander is included in
Appendix A.
In this assignment students are presented with a
script and the virtual world to go with it. They must devise
appropriate coverage for the action in this scene, capture images
for each camera setup, and indicate their coverage by lining the
script. The script for this scene is included in Appendix B. Sample
shots from this scene are included in Figure 14.
 
Figure 14
Conclusion
This project achieved its primary objectives: (1)
creating virtual worlds appropriate for teaching some of the basics
of visual grammar, and (2) designing presentations and student
exercises to take advantage of these virtual worlds. Work continues
on these simulations. We are attempting to add computer generated
guidance and critique which can be delivered while the student is
performing the exercise. The ability to give immediate feedback
appears to be just a bit beyond the capabilities of current
affordable technology, but this added capability, when available,
promises to make this simulation approach to teaching film and video
production even more effective. We are also developing more complex
simulations for use in advanced production classes.
Availability of Simulations
The Virtus "worlds" discussed above are available
for free from the author (michael_korpi@baylor.edu). |