Stephen A Kallis, Jr
1974
Focal Press
The recent increase in discussion of techniques for using electronic computers in motion-picture animation has been, in part, due to the greater sophistication of computer techniques coupled with the decreased costs of computer components. However, computer animation is becoming a nebulous concept that threatens to become virtually meaningless in the near future due to the misuse of both the term computer and the term animation.
In order to get a clear picture of what is happening, it is necessary to examine the various techniques that are all collectively termed computer animation. Also, it would not be improper to submit more precise terms for the different techniques.
Before we can discuss computer animation, however, it is important to define what is meant by computer. From a dictionary standpoint, anything that can be used for calculations is a computer; hence, a slide rule, a table of logarithms, or an adding machine could theoretically be called a computer. Here we use the term in its more restrictive sense - the sense in which it is used in the electronics industry.
There are two basic types of computer: the analog computer and the digital. In the former, input forces - or signals - are acted upon by certain conditions that affect the magnitude of the input(s). Analog computers may add, subtract, multiply, divide, differentiate and integrate inputs - and perform combinations of these functions - but always in a continuous way and on the basis of magnitude changes on continuous inputs and outputs.
By contrast, digital computers operate on discrete quantities. Basically, all digital computer operations are arithmetic (usually restricted to summing) and logical. Inputs to digital computers must, therefore, be in digital form or converted from analog signals to the numerical form.
Without digressing to discuss in detail the relative merits of each computer type, suffice it to say that any motion-picture process that is truly computerized must use one (or both) of these two computer types.
In one sense, digital computers lend themselves to motion-picture work somewhat more than analog computers, as both motion pictures and digital-computer operations are time-quantized phenomena. Therefore, we shall discuss the digital techniques first.
Digital computers are often employed in machine control functions. In that capacity, they are used to control both the position and the operation of machines such as drills and milling machines, making it possible, for instance, for a drill to make a highly complex pattern in a steel plate with a few input instructions. This process is known as numerical control in the production industry.
Having defined so far the concept of the computer, let us now give a brief definition of animation. Animation can be defined as the process by which static images or drawings (the artwork) are brought to life, i.e., the process by which they are provided with movement. The basic technique is the one by which series of individual, static images or drawings (the cels} with slight, controlled differences in their shape and/or placement, are photographed frame by frame with an appropriate motion-picture camera. Placement differences of the cels along the x,y,z axes (called east-west, north-south and zoom in animation parlance) are, of course, always relative to the camera; hence, a difference in placement of the camera is equivalent to a different placement of the cel. One or the other or a combination of both can be used, according to the type of movement that is wanted on the screen. (Rotational displacement of the artwork is another possibility.) These differences in shape and/or placement, when seen in rapid succession on the screen, produce the illusion of movement.
To discuss the basic types of techniques (of which all kinds of combinations are possible, of course), we shall follow this outline:
Computer-controlled camera and platen movements and functions
- indirect or off-line control
- direct or on-line control
Computer-controlled CRT display images
- static display images
- continuously variable (dynamic) display images
Following this classification we shall now consider the different possibilities computer animation offers.
COMPUTER CONTROLLED CAMERA AND PLATEN MOVEMENTS AND FUNCTIONS
Indirect or Off-Line Control
Numerical control functions can be adapted to animation equipment. What we believe is the pioneering studio to use this technique is Animated Productions, Inc., in New York. They are using a small Digital Equipment Corp. PDP-8 computer to generate control instructions to their animation stand (Fig. 1.1 ).
Fig 1.1: Off-line computer-controlled animation recording. Communication to the animation stand through an intermediate medium (in this case perforated paper tape) makes this an indirect or off-line form of computer control. This form is particularly suitable for situations in which one computer must service more than one stand.
In practice, the operation works this way: an operator determines what animation is required, i.e., panning, tilting, zooming, skip-framing, etc. The instructions are input to the computer via a Teletype, and the computer makes the necessary calculations.
After the calculations have been completed, the computer generates the necessary control instructions. In this particular system, the instructions are punched out onto a special paper tape.
The tape, in turn, is loaded onto a tape reader that is connected to the control motors and relays on a modified Oxberry animation stand. The reader-interface device then decodes the tape to electrical input signals to the Oxberry stand, which then performs the required operations.
Such a system is termed off-line because the computer does not control the stand directly; a direct hookup would be termed on-line.
What can be done with such a system? Obviously, such a setup permits pans and tilts (E-W and N-S movements, respectively) and rotation. Because of the nature of the computer, such movements can be made more realistic by permitting the computer to calculate the required acceleration and deceleration for such movements. Similarly, zooms can be made using the technique; and, indeed, even combinations of pans, tilts, rotation and zooms, all compensated.
The camera shutter is also under control, permitting precisely metered fades and dissolves. This is particularly important in the case of fades, since the characteristic curve of a film would cause a premature (and hence uneven) dropout at the lower end. Feeding information of the film's characteristic curve to the computer makes dropout compensation possible. Further, dissolves can be metered to permit unwavering image brightness during scene transitions.
With shutter control, too, other effects are possible. To perform a cyclic action in animation conventionally saves the artist's time and effort but requires as many cel changes as would be required for fresh art. With computer control of the shutter, frames may be skipped with excellent repeatability and without danger of errors, as in human operation. In the simplest cases, the total number of cel changes would be one less than the number of cels required for one single, complete cycle. With proper computer control of both foreground and background, this precision skip-framing can be used even when background and foreground are moving in relation to each other. This can increase efficiency tremendously. (A 12-cel cycle could be filmed in 12 passes of the film, no matter how long the duration of the scene.)
First, all No. 1 frames of each cycle are exposed by the skip-frame technique, the shutter remaining closed for all the remaining frames of each cycle. The film is then rewound and all No. 2 frames are exposed in the same way, and so forth until all frames of all 12 cycles are exposed. Shutter function and rewind with closed shutter are, of course, fully programmed in advance. If the forward and backward speeds of the camera, and hence both exposure times, are exactly alike, half of these passes can even be eliminated by adequately skip-framing on the back run. This would save still more time - as opposed to, say, 239 artwork changes for a 10-s scene with 20 complete 12-frame cycles.
At this point, it is important to point out that this technique uses ordinary animation cels and techniques for the artwork. For this reason, this form of computer animation deserves a more precise name. We believe the most descriptive and accurate would be computer-controlled animation recording (computer recording).
One extremely important aspect of computer-controlled animation recording is that it is completely repeatable. Should the original footage be destroyed or damaged, it is easy to set up the artwork and reshoot the scene precisely, something that is practically impossible to do by conventional means. Further, because of the control features of this form of animation, it is possible to rack over the camera (disconnect it from the system) and perform a dry run at maximum speed if this is desired.
In the off-line system, the control tapes are often saved and utilized again rather than tying up the computer. For a studio with several animation stands, this would permit several stands to use the computer's services without interfering with each other.
Fig 1.2: On-line computer-controlled animation recording. Direct communication to the animation stand via interfacing makes this an on-line form of computer control. With the addition of extra memory in the computer, this form of controlseems especially suited for the production of long and complex animation sequences.
Direct or 'on-line' control
Basically, on-line computer-controlled animation recording is performed in much the same manner as off-line control except that the computer is connected directly to the animation stand through appropriate interfacing (Fig. 1.2). Such an arrangement has both advantages and drawbacks.
The basic operating principles of computer-controlled animation recording remain the same whether the system is off-line or on-line, so there is no need to repeat the obvious advantages of this technique over that of conventional animation. Obviously, all the control signals punched into an off-line tape are contained within the computer, so if they could be transferred directly to the animation stand without having to go through the additional steps, the time factor would be reduced considerably. Further, the interfacing requirements would be simpler than would be necessary with a tape reader.
Against this, an on-line system ties up the computer, which could be doing other things. In a studio with only one computer and only one animation stand, this would be little or no handicap, but in a studio with multiple animation stands, a financial assessment would have to be made between the cost of additional computer systems and the use of a single computer off-line; though with the trend to lower computer costs, this might soon be a sterile point of debate.
Thus, the verdict is not in yet - if it ever will be - as to whether the on-line or off-line technique is superior for computer-controlled animation. However, either method will permit the production of precise animation with orders-of-magnitude decreases in production time.
COMPUTER-CONTROLLED CRT DISPLAY IMAGES
Static display images
Beyond the augmenting of conventional animation techniques, the second major form of computer animation uses the computer to create the artwork as well as to record it. For a number of years computers have been coupled to visual displays, usually in the form of cathode-ray tubes (CRT), though normally not employing raster scan. By means of appropriate program entries, a computer can display a plot on a scope face of assigned numerical co-ordinates. Each numerical position then can form a dot. In some CRT systems, the dots are strung together to form straight or curved lines, each defined mathematically as the locus of point co-ordinates comprising its dots. In another vector-oriented method, the electron beam is moved incrementally by coordinate, vector and scalar entries.
There are two main classes of CRT used for this purpose. These are known as refresh scopes and storage scopes. A refresh scope is most like the conventional CRT in which a formed image has to be retraced (or refreshed) continuously to remain visible, analogous to the field renewal of a conventional television receiver. By contrast, a storage scope is so constructed that a phosphor activated by the tracing electron beam continues to glow for an extended period (perhaps as long as 15 min). The storage scope is particularly well suited for a complex vector-oriented image from a complex series of calculations (Fig. 1.3).
Fig 1.3:Mechanical plot of a complex vector-oriented image of a warped plane in perspective. This presentation is identical to the type seen ona storage CRT.
Either type of CRT can be used for this form of computer animation. Through more sophisticated programming hidden-line algorithms can be formulated to inhibit the tracing of lines that ordinarily would be hidden behind a portion of a figure, thus minimizing visual confusion.
By use of these techniques it is possible to place on a scope face a series of similar figures, serially. By taking pictures of successive images by a stop-motion camera setup, it is possible to project animated film.
Many computer enthusiasts have experimented with this approach. In the similar setups, a camera is slaved to the computer output, usually by a solenoid. As each computer-generated frame is presented on the CRT, a signal is sent to the camera, which records the image (Fig. 1.4).
Fig 1.4: Typical static computer generated animation imaging set-up. In some of the more primitive forms of this technique the instructions may be entered in the form of punched cards. This form is primarily used for production of precision engineering and scientific solutions to problems for studies and classroom work.
More elaborate devices use a computer-output microfilm (COM) device to achieve this. In such a device, a high-resolution CRT is focused on microfilm, and the graphical computer output is recorded. A COM was originally used as a convenient way of recording permanently graphical computer outputs, but the device lends itself to this form of animation.
Unlike computer-controlled animation recording, this technique requires little mechanical control of camera placement, etc., save film advance, but it does need rather elaborate programing to generate the figures. For this reason this approach is termed static computer-generated animation imaging (or static computer imaging).
Static computer-generated animation imaging is perhaps the most precise form of animation possible. Every element of the process is under complete control and can be repeated exactly as is true in computer-controlled animation recording; however, in addition, the artwork is virtually indestructible, since it is in the form of mathematical entries.
There is a corresponding disadvantage with computer-generated animation imaging. The degree of complexity of writing equations for figures increases greatly when the figures become more complicated than simple geometrical figures. Further, some methods of graphical plotting require extremely long times for complex figures, making exposure-time requirements that would be intolerable for many camera systems.
For simple figures, however, such as the dynamic solution of problems of moving bodies in physics problems, this technique is unsurpassed. It may prove to be the foundation for many teaching programs in the future.
Continuously variable display images
In the context of this chapter, an image is continuous if its refresh rate is faster than the frame or field rate of the picture taking device.
Another method of generating images for motion pictures utilizes an analog computer. While detailed approaches vary, the basics of this type of image generation are due to the nature of analog computers. A signal being input into an analog circuit can be altered by the analog-circuit elements, changing its form or the shape displayed.
To clarify this technique, let us take the simplest example we can think of: let us suppose that we feed 60-Hz signals of equal magnitude into the horizontal and vertical deflection elements of a CRT. The image of the electron beam trace that we would see on the tube face, in these conditions, would be a straight line at a 45° angle, assuming that the two 60-Hz signals are exactly in phase. But if we add an analog circuit that can shift the phase of either the horizontal or the vertical signal with relation to the other, a twist of the control would alter the image from a line through ellipses to a circle, and continuing, back through ellipses to a straight line that is at right angles to the original line.
Analog devices that can alter frequency, wave-shape, amplitude and similar variables can manipulate the image on a CRT; this is the basis of the third type of computer animation. By using complex analog computers, highly complicated artwork can be generated on a CRT face (Fig. 1.5).
Fig 1.5: Images generated by analog computer elements. In this arrangement, a graphical pattern has been used as the input element and the system is being used to distort it (Courtesy of Computer Image Corp.)
Figure 1.6 is a block diagram of an equipment setup for one form of dynamic computer imaging. A specially interesting element is the so-called pattern scanner whose function is that of a camera. In this sort of arrangement, an initial pattern (such as a title) is scanned by the pattern scanner and is converted into analog signals in approximately the same way that a television camera scans a scene. The images thus scanned can then be displayed on a CRT screen and manipulated through appropriate circuitry. Two devices utilizing this principle are presently employed by Computer Image Corp. in Colorado and elsewhere.
Fig 1.6: One form of dynamic computer-generated animation imaging. This form is suitable for manipulating art work images and is especially useful in making titles for both motion picture features and television. Other forms do not require the input of an external image but generate the 'artwork' from circuits alone.
In theory, such an analog arrangement can manipulate any input electronic signal. One method of producing an original signal to be manipulated is by means of an image-scanning device; another is by means of circuits that generate basic signal forms such as the straight line and circle previously cited. In either system approach, the image thus formed is usually photographed in real time, that is, at 24 frames/s. Because of this, and because the classical concept of animation implies a compression of time quantization, this third type of computer animation may be called dynamic computer-generated animation imaging (dynamic computer imaging).
Dynamic computer imaging does have certain inherent qualities not available with the other techniques. Because of its nature, images can be examined in real time before the picture is taken, thus permitting the photographer to see a perfect dry run. Corrections are, of course, performed electronically and can conceivably be achieved by the manipulation of a single control.
Because of the purely electronic manipulations involved with this technique, preparation of artwork is kept at a minimum and is certainly far more rapid than the programming required for static computer imaging.
Both static computer imaging and dynamic computer imaging have enjoyed commercial success, and as a result there are occasional discussions as to which of the two is the better method. It would be unwise to make a value judgment during this early stage of development of the two techniques. At any rate, the answer to this question is perhaps as meaningful as the answer to the question whether television or motion pictures is better. There is no legitimate answer because the techniques are rather more complementary than competitive, and time alone will tell whether the two techniques will combine to some form that takes the feature of both or whether each will develop along its own lines. They might even be combined with computer-controlled recording.