Towards
an Integrated Computer Art System
|
||||
Published
in Computers in Art, Design and Animation, Springer Verlag, pp.
643 - 652, 3,300 words |
Computer
Art Media FIGURE 1.
Primitives in Macdraw.
FIGURE 2.
Basic operations.
FIGURE 3. Classical and recursive geometries. The distinction between classical and recursive geometries lies in the computing techniques behind them. In a recursive geometry the positioning of gometrical elements such as lines or motifs is the result of successively using the output of one calculation as the input to the next. The classical geometries, while often using some repetitive technique, do not base a given calculation on the previous one. This is a slight simplification, and in fact a recursive geometry can be created without using the programming technique called recursion. To illustrate the idea of a recursive geometry, consider the diagrams in Fig. 4, which show a line segment (initiator) being replaced by a shape consisting of eight line segments (generator). On each "recursion" or generation in the production of the image, each existing line segment (initiator) is replaced by the generator, scaled down to fit.
FIGURE 4. Recursive (fractal) trees. There is not space here to describe in detail all these geometries and algorithms or their uses to a computer artist, but I will discuss a few. Nets, bands and tessellations are very important in working with pattern and are well described in Macgregor and Watt (1984). Figure 5 shows a simple net of motifs (motifs placed on a regular grid), while Fig. 6 shows tessellations produced from an interactive program called Tessellator (Addison-Wesley). FIGURE 5. A net of motifs. FIGURE 6. Simple tessellations. Non-recursive functions have been important for a long time in computer art. These functions are used to control the x and y position of some primitive (often just a dot or line) and the output is in effect a type of graph of the function. Franke (1971) shows many such images, including his own work, while Leavitt (1976) again shows the work of a variety of computer artists who have used this technique. Figure 7 shows output from a function designed by computer artist Joseph Jacobson (1982). FIGURE 7. Output from a mathematical function. FIGURE 8. The Mandelbrot set. Parametric curves, such as the B-spline, are of importance in computeraided design for automotive and aerospace design and allow the description of curved lines and surfaces from a few control points. These can be very useful to the computer artist (in particular the computer sculptor) and have already been exploited in the Rodin system. The recursive geometries offer the artist imageries and techniques quite unique to the computer, such as fractals and graftals. Figure 8 shows the Mandelbrot set, and Figure 9 shows a simple tree generated from a set of rules known as a grammar and described by Alvy Ray Smith (1984) as a "graftal" -something like a fractal but not quite the same. FIGURE 9. A graftal tree. I see the collection of geometries and techniques listed above as making up a toolkit for the computer artist, a toolkit which is of course open-ended - new developments will continually add to it. The power of these techniques is that they can be combined in ways unique to a particular artist's way of thinking or way of exploring. Output from different techniques can be "matted" together, that is, just overlaid in different ways, or they can be combined in more fundamental ways, where output from one function, for example, may modulate the output from another. Some of the elementary classical geometries are provided within paint and drafting systems, but as soon as the user attempts to create more complex geometrical structures it becomes very difficult. Even the simplest kind of pattern such as a half-drop often has to be created by copying the motif individually, which is absurd when the programming of such a geometry is a trivial exercise. In order to explore the more complex geometries, such as fractals, the artist is forced firstly to research the techniques and secondly to write programs from scratch to implement them. If the artist wishes to combine various geometries, then again they have usually to be programmed from scratch. The Importance of Computer Geometries The computer offers firstly free-hand or interactive methods of working, which allow relatively traditional approaches to image generation to be used, and secondly programming approaches. If the artist is going to use programming, then computer geometries, both classical and recursive, open up vast territories for exploration. It is easy to dismiss these techniques as pattern making of a sophisticated kind, but this does not do them justice. Firstly, the image generated from a mathematically or geometrically based program does not have to be the end product: it can be used as a starting point for interactive techniques as in a paint system. Figure 10 illustrates this with an image that started as the fractal trees of Fig. 4.
Secondly,
mathematical and geometrical patterns have a profundity about them due
to the fact that they describe laws of nature. This, of course, is the
attitude of the Islamic artist, who sees the geometries in terms of cosmology
and astrology. Richard Voss, in an article on fractals (Voss, 1985), quotes Galileo:
Proposals for an Integrated Computer Art System If one accepts that computer art needs both the interactive "arbitrary" approach and the algorithmic computer geometries, then what kind of system allows the artist to exploit and mix these methods? The answer, at present, is that no such system exists and that many existing systems work against this approach, either by being "closed" paint or drafting packages which cannot be extended with the various geometries or by being programming environments with no interactive components. A system that offered both approaches could be called an integrated computer art system, or ICAS for short. The system that comes closest to being an ICAS is the Juno system built by Greg Nelson (1985) as a prototype. This is an interactive drafting system with a scripted component where geometrical entities can be specified using a simple language of "constraints." The use of constraints represents an attempt to develop a more powerful programming language than the usual languages such as BASIC, Fortran and Pascal. I refer to this as "scripting at a higher level." An ICAS should incorporate all the geometries so far discussed and allow interactive synthesis at many different levels (solid modelling, drafting and painting). Analysis, in the form of image processing, or other methods of taking data from the real world must be catered for. An ICAS should offer: Interactive and scripted components, including solid modelling, drafrting, and painting. Image-grabbing and image analysis. Graphics hardware for bit-blitting and similar operations (moving chunks of picture). A general programming environment. Scripting at a higher level such as constraints for classical and recursive geometries. Frame-buffer output and vector outut to plotters. In working with an integrated system one could just use the paint system facilities if desired, but a more usual route would be to start at a more modelled level of construction. One could, for example, script a geometrical algorithm which would allow many possible types of output and rendered image; this script would act as the most abstract description of a series of images. At a lower level one might take the output of the script (for example a sequence of vectors) and manipulate these with drafting techniques. Finally, a single image could be created by taking the output from the drafting part of the system into the painting part, where the image can be manipulated on a pixel basis - for example, using flood-filling and the creation or addition of "biomorphic" elements, as in the Mandelbrot set example of Fig. 11. Other routes would include three-dimensional modelling, the use of standard visual realism techniques and the integration of different elements within the geometrical toolkit. Yet another starting point could be with a frame-grabbed image and the manipulation of this data in various ways. Analysis of such an image could be the start of a Markov chain approach to generating imagery or even just the creation of texture for certain purposes. An interesting technique devied by Brian Reffin-Smith while at the Royal College of Art was to calculate the angle of a line segment based on the pixel value in a frame-grabbed image and output the results onto a plotter. I have concentrated so far on the computer artist rather than the designer or graphic designer. I believe, however, that the untrammelled explorations of the computer artist will have an impact on the vocabulary, certainly of the graphic designer, who will make increasing use of the computer even for the most trivial and conventional of briefs, and also on the product designer and architect. If a revival of interest in ornament in architechture and design should take place, then the computer will be at its centre and should have a profound inpact on its nature. Algorithmic synthesis from primitives and the elements of the geometrical toolkit could bring about a new interest in decoration and ornament, but with a contemporary style. Conclusions The computer offers interactive or free-hand techniques and also programmed approaches which allow the exploration of more sophisticated computer geometries. An integrated computer art system would incorporate features from the major interactive systems with a scripted or programmed component, allowing the artist to explore both traditional geometries and those which arise from recent research in mathematics, physics and biology. I believe that widespread use of such systems would make commonplace a new visual vocabulary, which would have an impact at the least on areas traditionally involved with pattern-making. References Critchlow, K. (1976) Islamic Patterns. Thames & Hudson. Dixon, R. (1983) "Geometry Comes Up to Date", New Scientist 98(1356), May 5:302-305. Franke, H. (1971) "Computer GraphicsåComputerArt" Phaidon. Jacobson, J. ( 1982) "Analytical Computer Art," 1982 IEEE Symposium on Small Computers in the Arts, pp. 47- 55. Lansdown, J. ( 1978) "The Computer in Choreography," IEEE Computer, August, pp.l9-30. Lansdown, J. ( 1980) "Is the Computer a Tool?: The Question in an Art Context." in B. Sundin, ed., Is the Computer a Tool? Almqvist & Wiksell, Stockholm. Leavitt, R. (1976) Artist and Computer. Harmony Press. Macgregor, J. and A. Watt (1984) TheArt of Microcomputer Graphics. AddisonWesley, Reading, Mass. Mandelbrot, B.B. (1982) Fractals: Form, Chance and Dimension. Freeman, San Francisco. Nahas, M. and H. Huitric (1982) "Computer Painting with Rodin," 1982 IEEE Symposium on Small Computers in the Arts, pp. 95-103. Nelson, G. ( 1985) "Juno, a Constraint-based Graphics System," Computer Graphics 19(3):235-243. Smith, A.L. ( 1984) "Plants, Fractals and Formal Languages," SIGGRAPH 84, pp. 1-10. Voss, R.F. (1985) "Random Fractal Forgeries." In R. A. Earnshaw, ed., Fundamental Algorithms for Computer Graphics," NATO ASI Series. Springer-Verlag, New York. Wilson, S. (1983) "Artificial Intelligence in the Arts," Leonardo 16(1):15-20.
|