Introduction - Computer Animation: Algorithms and Techniques - page 3

While perception of motion addresses the lower limits for establishing the perception of continuous
imagery, there are also upper limits to what the eye can perceive. The receptors in the eye continually
sample light in the environment. The limitations on motion perception are determined, in part, by the
reaction time of those sensors and by other mechanical limitations such as blinking and tracking. If an
object moves too quickly with respect to the viewer, then the receptors in the eye will not be able to
respond fast enough for the brain to distinguish sharply defined, individual detail; motion blur results
[ 11 ]. In a sequence of still images, motion blur is produced by a combination of the object's speed and
the time interval over which the scene is sampled. In a still camera, a fast-moving object will not blur if
the shutter speed is fast enough relative to the object's speed. In computer graphics, motion blur will
never result if the scene is sampled at a precise instant in time; to compute motion blur, the scene needs
to be sampled over an interval of time or manipulated to appear as though it were [ 21 ] [ 32 ]. (See
Appendix A.3 for a discussion of motion blur calculations.) If motion blur is not calculated, then
images of a fast-moving object can appear disjointed, similar to viewing live action under the effects
of a strobe light. This effect is often referred to as strobing . In hand-drawn animation, fast-moving
objects are typically stretched in the direction of travel so that the object's images in adjacent frames
overlap [ 49 ], reducing the strobing effect.
As reflected in the previous discussion, there are actually two rates of concern. One is the playback
or refresh rate—the number of images per second displayed in the viewing process. The other is the
sampling or update rate—the number of different images that occur per second. The playback rate is
the rate related to flicker. The sampling rate determines how jerky the motion appears. For example, a
television signal conforming to the National Television Standards Committee (NTSC) format displays
images at a rate of roughly 30 frames per second (fps), 2 but because it is interlaced , 3

fields are played
at 60 frames per second to prevent flicker under normal viewing conditions [ 34 ]. In some programs
(e.g., some Saturday morning cartoons) there may be only six different images per second, with each
image repeatedly displayed five times. Often, lip-sync animation is drawn on twos (every other frame)
because drawing it on ones (animating it in every frame) makes it appear too hectic. Film is typically
shown in movie theatres at playback rates of 24 fps (in the United States) but, to reduce the flicker,
each frame is actually displayed twice ( double-shuttered ) to obtain an effective refresh rate of 48 fps.
On the other hand, because an NTSC television signal is interlaced, smoother motion can be produced
by sampling the scene every 60th of a second even though the complete frames are only played back
at 30 fps [ 8 ] . Computer displays are typically progressive scan ( noninterlaced ) devices with refresh
rates above 70 fps [ 34 ] . See Appendix B.10 for some details concerning various film and video
formats.
The display and perception of animation using a sequence of still images imposes certain require-
ments on how those images are computed and played effectively. Understanding the operation, limits,
and trade-offs of the human visual system are essential when making intelligent decisions about design-
ing any type of visual and auditory content, including computer animation.

2 More accurately, the format for broadcast television system, established by the NTSC, specifies a frame rate of 29.97
fps [ 29 ].
3 An interlaced display is one in which a frame is divided into two fields . Each field consists of odd or even numbered scan-
lines. The odd and even fields are displayed in alternate scans of the display device [ 8 ].