Coda: Provocations on Filmic Retrieval, Hunting, Meandering, and Browsing – Video Structure Meaning

55
row window of entropy necessary for maintaining the illusion of motion; too much entropy and the
document loses coherence, while too little entropy and the document no longer engages the viewer.
00:01:02:29 00:01:02:30
00:01:03:15
Synthetic Frame
Synthetic Frame
00:01:03:16
Figure 2.15: Synthetic frames showing small frame to frame variation (above) and large (below).
2.5 STORY FIVER: WHAT MAKES A MOVIE AND WHY DOES IT MATTER?
56 2. FIVE STORIES TO A MODEL OF VIDEO STRUCTURE
We need a little more though. e illusion of motion is normally
brought about by the slight changes in data from frame to frame when
played back at the intended or nominal speed of the medium. A viewer
of a collection of random photographs could arrange a set of prints
or digital les and allot a set time period for viewing each image and
set an order in which they would be viewed, but this would not nec-
essarily present any perception of motion, nor would it necessarily be
considered a representation of motion. It would be, essentially, a slide
show; it might have thematic coherence, yet would not be a moving
image document.
For another case to consider, we turn to a recent development
in video to nd a transitional case—the Ken Burns eect. Documen-
tarian Ken Burns developed a technique by which Action is given
to still photographs by slowly zooming in on subjects of interest and
panning from one subject to another.” e illusion of motion is gener-
ated by moving the camera (or software version of a camera) over the
image, thus producing a set of frames that have the sort of dierence
between any two consecutive frames we discussed above. e image
on the screen, the stimulus set to the eyes of a viewer, is changing at
a standard rate; the illusion of motion though is motion of the still
photograph rather than of the objects in front of the original camera.
Here a sample of frames from two seconds of panning to the left across
an image of a city street.
is is not necessarily a cheat in terms of message making or
story telling and the eect does depend on the same persistence of vi-
sion that seems to account for what would normally be called a movie,
yet there is no illusion of motion in the ordinary sense of some objects
moving against a static backdrop and with regard to one another. We
are speaking here of message making, of a lmmaker coding a mes-
sage; as Blair [Hayes too sort of ] suggests, the lmmaker dances with
the viewer, making assumptions about the viewers decoding abilities.
Persistence of vision sets limits on coding practices; it frames the rate of change in the visual data
stream at playback. Too little change from frame to frame and the viewer perceives no motion; too
much change from frame to frame and the ability to merge the data is lost.
Any single pixel address within a frame is comprised of four values: red, green, blue, and
opacity—RGBA or RGBα. For any pair of frames two additional values are added to the pixel
57
address data: directionality and magnitude. ese form a vector describing the amount of change
over time; in a movie this period is now ordinarily 1/30th of a second.
So what? We assert that the meaning of a video document—of any document—is its func-
tionality for the author and its functionality for any viewer. Any document is structured data (coded
data). Movies present movement. In order to analyze movies to understand how they are coded to
generate meaning and, at the same time, to develop methods of categorizing movies based on their
coding structures—what might be called ngerprinting—we need to be able to describe movement
in rigorous terms. We need to be able to describe and compare sorts of motion without losing sight
of the motion. In order to facilitate retrieval and understanding of moving image texts we have to
come to grips with the precarious balance between stillness and movement we encounter at every
step. Ultimately, it is the transit from one still image to another which has defeated every attempt
to freeze the lm-text as it turns into something else—a still photograph if you stop the machine,
or any number of quasi-cinematic representations if you slow it down or speed it up. Yet, it remains,
in Bellours term, unattainable in motion—“le texte introuvable.”
2.5 STORY FIVER: WHAT MAKES A MOVIE AND WHY DOES IT MATTER?