by Dan Collins,
Associate Professor of Art, Arizona State University
This paper was first given at the 6th Biennial Symposium on Art and Technology, Connecticut College, February 27 - March 2, 1997
Digital sculpture draws upon recent advances in data acquisition techniques, computer visualization, and rapid prototyping technologies. It utilizes the unique virtual space of the computer to pre-visualize form, to enable extraordinarily sophisticated formal innovations, to design at heretofore unmanageable scales with technical accuracy, and to produce objects impossible to create with the human hand. It opens a floodgate of questions regarding the use and future use of a technology that is predicated upon a "rapid response" to the needs of a culture.
Before addressing the larger question of "how to become a better tool user," let me address the convergence of technologies behind what I am calling "digital sculpture."
Three domains must be understood and mastered by the digital sculptor: Data Acquisition (input technologies); Computer Aided Design, modeling, and visualization (CAD), and Computer-Aided Manufacturing (CAM).
Data Acquisition: From Microscopes to Satellites
The first domain, Data Acquisition, covers a wide array of systems operating at a full spectrum of scales. Virtually anything that can be mapped in 3D can be rendered as a digital sculpture. Extremely tiny objects--e.g., blood cells, crystalline structures, larger molecules at nano scales--can be rendered as three-dimensional models using tools such as a Scanning Probe Microscope. Objects are measured in nanometers--units a billionth of a meter long. Slightly cruder in its resolution is the Confocal Microscope which is adept at rendering three-dimensional objects at micro scales. Models are measured in microns--units a millionth of a meter long. At larger scales, objects several millimeters across can be digitized with manual probes or optically scanned using 3D laser scanners. Depending on the lens configuration, resolutions as fine as .125 mm can be achieved. A full body scanner has been developed by Cyberware of Montery California that records a million data points spaced at 3 mm intervals. Medical diagnostic technologies such as MRI, CT scanners, and 3D Ultrasound give three-dimensional representations of internal morphology. In combination with surface scanners like the Cyberware machine, these technologies enable a new kind of "figurative sculpture" that reveals both internal and external anatomy. At very large scales, terrestial maps or extraterrestrial planetary surface data are delivered by satellites equipped with systems such as Synthetic Aperture Radar.
Computer Visualization: Modeling the Data
Much as the drawings of Leonardo, Vesalius, and Alberti achieved a synthesis among techniques of observation, systems of measurement, and the pure pleasure of representational drawing and painting, recent work in computer visualization provides a link between contemporary aesthetics and science. In many respects, computer visualization is where a quantifiable description of the world--the traditional domain of the physical and life sciences--finds new expression and meaning in enhanced methods of representation--the traditional domain of the arts and humanities.
"Computer visualization" is an area familiar to anyone who has struggled with the learning curves typical of 3D modeling packages. At its most arcane, the task of translating quantifiable information into pictures falls to mathematicians, programmers, and software engineers trained in Computer Aided Geometric Design (CAGD) and Computer Graphics. Most sculptors do not have the requisite technical skills to develop the algorithms or programs that are necessary for translating abstract spatial concepts into 3D models on the screen. However, artists comfortable with working in three dimensions will find that their skills at visualization, "imaging", and conceptualization are highly valued in technical circles. As one becomes more adept at the protocols of a 3D digital universe, a healthy appreciation of the complexity, logic--and especially the elegance--of software and hardware design will grow.
We are seeing increasing numbers of successful collaborations between individuals with technical competencies and those with an eye for the aesthetic or the conceptual. Major conferences in computer graphics such as SIGGRAPH highlight the latest breakthroughs in art, industry, entertainment, and theoretical computation. The entertainment industry drives much of the technology and provides opportunities, challenges, and applications for the development of complex algorithms and new programming protocols. The special effects wizardry achieved by Steve Jobs's company, Pixar, which produced the feature length animation, Toy Story, and the legendary work of George Lucas and his company Industrial Light and Magic (which has just re-released the epic trilogy, Starwars) are just two examples of companies that are built on the premise of collaborative enterprise between technologists and artists. Apart from the kinds of spectacles possible with the deep pockets of the entertainment business, there are some unique, innovative projects that bring together the best the arts and sciences have to offer. The CAVE project (Cave Automatic Virtual Environment) at the Visualization laboratory at the University of Illinois at Chicago is perhaps one of the more innovative examples of collaborative enterprise integrating high end graphics into an immersive 3D audio/video experience. Viewers wear special glasses utilizing "shuttering" technology (LCD stereo shutter glasses are used to separate the alternating video fields being received by the eyes), and move through a space "guided" by a leader wearing a full VR headset. The correct perspective and stereo projects of the environment are updated, and the image moves with and surrounds the viewer. Content explored has included everything from the visualization of natural phenomena to scientific data to "fine art" immersive environments.
Many related areas converge in computer visualization. Computer graphics, scientific visualization of all kinds, 3D Modeling, 3D Archive/Database development, 3D Pattern Recognition (requiring artificial intelligence expertise) are just some of the areas of interest to the digital sculptor.
A set of tools that can accomodate the diversity of data generated by the range of data input devices ennumerated above is essential. In the PRISM lab--a research unit devoted to visualizaton and prototyping at Arizona State University--data gathered from machines as diverse as Scanning Electron Microscopes, MRI machines, and 3D laser scanners has been successfully translated into solid models. We are currently working on a method to translate satellite data into 3D models using Geographic Information Systems software called ArcView.
Form Realization: 3D Printing and Rapid Prototyping
At the heart of the process of digital sculpture is the desire to translate three-dimensional objects designed in the virtual space of the computer into actual three-dimensions--a process generally known in industry as "rapid prototyping." Rapid computer-based prototyping--more precisely called "layered manufacturing"--is a relatively new field that is gaining increasing acceptance in fields as diverse as medicine, aerospace, industrial engineering, and sculpture. Complex three-dimensional objects such as human prostheses, molded tooling, or aerospace parts are produced by linking Computer Aided Design (CAD) products with various rapid prototyping (RP) systems. It is also an ideal tool for concept modeling as RP systems can provide CAD-equipped design engineers (and digital sculptors!) with a physical model of a proposed design that they can touch, examine in detail, and ship to others for inspection and review.
There are numerous automated devices that can translate 3D CAD models into tangible form. Most common are the Computer numerically controlled (CNC) machines that have been in service since the 1970s. CNC milling enables computer controlled cutting on multiple axes in a variety of materials. Earlier electromechanical machines--both analog and digital--capable of cutting relatively complex solids were introduced just after World War II. Other machines regularly used in heavy industry include computer controlled plasma and laser cutters, electro discharge machining (EDM) systems, automated hi-pressure waterjet cutters, and various sand and glass bead blasting technologies, to name a few. Most of these tools have unique control protocols that do not port easily from one machine to another.
An entirely new category of manufacturing processes are classified under the rubric of "layered object manufacturing." All of these new systems--irrespective of the methods with which the final part is made--require "solid models" that are subsequently "sliced" (digitally) into multiple horizontal "layers" that can be produced in sequence by a given process. Some machines cut individual sheets of paper (with razor blades or lasers); others deposit thin semi-liquid threads of wax or plastic; still others use lasers to harden liquid resins, fuse synthetic "flours", or, in the most exotic processes, cause precise chemical reactions to occur at given points within a gas-filled chamber (only at very small scales).
Within the field of layered object manufacturing there is a rapidly growing number of competing technologies ranging from small desk-top "3D printers" and cutter/plotter devices to room-sized "rapid prototyping centers." Each technology has its advantages and disadvantages. As might be expected, the desk-top units are relatively inexpensive. A computer aided plotter device that cuts individual sheets of adhesive backed craft paper machine can be purchased for less than $7500.00. At the other extreme, a laser "sintering" (fusing) machine can cost nearly half a million dollars. In these high end machines, the hard-copy output is remarkable for both the fineness of surface detail and their robust structural integrity. Advances in the range of materials available for prototyping permit particle sizes of as small as 40 microns (e.g. DTM's"True Form") which can be fused for fine feature definitions down to .004 inches resolution. Middle range desktop machines such as the Stratasys/Genisys thermoplastic extrusion system ($50-60,000.00) can achieve resolutions of .020 inches. A short list of the companies that manufacture RP devices include the following: 3D Systems pioneered the so-called "stereolithographic" systems (SLA) in which a bath of photosensitive resin is hardened by a laser. Stratasys specializes in a process known as Fused Deposition Modeling which involves the extrusion and precise layup of a continuous thread of hot thermoplastic. Cubital uses ultraviolet light to harden full layers of lightsensitive resin. Ballistic Particle Machining (BPM) uses what is essentially an inkjet technology to build up layers of hot synthetic wax.
Because of the incredible cost savings that can be achieved by shortening the production cycle in industry, there is substantial interest in all kinds of rapid prototyping. Many of the initial problems associated with rapid prototyping--toxicity, lack of durability, problems in translation to metal or cold casting methods--have been solved. A handful of companies have already perfected methods for directly producing production level tooling (for injection molding machines, in particular) using RP technologies.
Goals for Digital Sculpture
As artists and designers, we are at the beginning of a revolution in form making that promises to transform radically the way in which we conceive, design, and create the objects in our environment--and our environment itself. Software and hardware is already available that can address a full range of materials from micro to macro scales.
In my own work, I have experimented with a number of methods for "intervening" in the process of form making. An artist can significantly alter outcomes in any of the three domains sketched above.
At the level of "Data Acquisition," for example, changes in the character or velocity of the object being scanned can radically alter its final form. It is now quite common for a traditionally trained artist to take their hand-made maquette (a small model of the desired finished sculpture) to a service bureau to be "digitized" using a 3D laser scanner or a single point digitizer. It is less common to see an artist deliberately impede the scanner's ability to faithfully record surface data. In my work, I have experimented extensively with moving the scanned object in relationship to the moving laser. This technique leads to what I call a three-dimensional blur of the data set. I have also worked with materials of different reflectivities and densities to see how they would alter the final digital model.
At the level of "Computer Visualization" the goal has been to radically alter the form of the base object by introducing new algorithms into the data sets that describe the objects. It is possible to introduce "anamorphic" distortions that appear to optically compress or expand a form depending on one's vantage point. To date, I have achieved anamorphic effects in two ways: 1) by simply compressing the radius dimension of a given object producing, in effect, a "stretched" form, and 2) by "morphing" the object with a known volume that mimics the "cone of vision." Ideally, a more accurate method of distorting the data set--one that is rooted in the physics of the eye--will be forthcoming.
A second goal in Computer Visualization has been to merge diverse data sets. Most high-end 3D modeling packages will support various Boolean-type operations for merging data in various ways. Some operations constitute relatively simple "unions" of disparate data into one discrete form. Other modifications include introducing dynamic information into the data such as sine waves or twist algorithms. Still other operations "interpolate" between two or more data sets to achieve what computer graphics artists call "morphing." The popular image of Michael Jackson "morphing" into various other people and animals is a good example--although this would be strictly two-dimensional information. One need not limit the field to living or organic objects. Consider what would happen if a human hand were "morphed" with an everyday object like a pair of scissors. By controlling the relative percentage of influence one data set or the other had on the final outcome, one could either have a sculpture that was "scissor-like hands" or "hand-like scissors." It is up to the artist/designer to determine the correspondences and differences that would be emphasized and / or de-emphasized. While the resulting sculptures would be certainly visually exciting, there are implications in this kind of research for scientific fields as diverse as cultural anthropology, genetic engineering, and ergonometrics, as well as practical applications for craft operations such as tool manufacture and furniture design. Here unfettered visual research takes the lead in pushing the envelope of what is possible in terms of physical form.
A third goal in Computer Visualization has been to understand the discrete differences of "hand-made" versus "machine-made" objects. What are the opportunities for "hands-on" modeling within the space of computer? One idea involves the manipulation of the data directly in the VR (virtual reality) environment of the computer. The goal here would be to produce an interface that is essentially transparent. While a mouse or a graphics tablet can afford a certain amount of freedom in the design process, what several research teams are exploring are VR based systems that utilize "data gloves" or other more ergonometrically responsive input methods (such as spatial positioning sensors or sound translators) that permit interaction with the virtual data on screen in real time. The HERA group (an acronym standing for Hand Eye Relational Analysis) at Arizona State University has made exciting progress in this area. They can currently manipulate a free-form solid using inputs from a conventional data glove whose position is tracked spatially by infrared sensors. This requires lots of computing power--and models that respond in extremely subtle ways to touch, physical pressure, or other human factors.
At the level of Form Realization or output, little research has been done into the idea of physically intervening into the rapid prototyping process itself. While one already has a full palette of choices with respect to the final outcome of a computer-designed object via rapid prototyping (e.g., choice of machine, material, scale, etc.), it is exciting to consider ways in which to physicaly intervene into the "build process" in a manner that will lead to predictable (or at least visually interesting!) outcomes. For example, with the Stratsys Genisys machine that we use in the PRISM lab (a thermoplastic extruding device), we'd like to attempt to introduce foreign matter--tiny rods, bearings, functioning electronics--into the material as it is being extruded. One could exploit the potential of the extruded thermoplastic as simply a method of holding a variety of diverse elements in a particular relationship. Charles Thomas at the University of Utah has successfully introduced "foreign" elements in the form of axles and a motor that were built into a rapid prototype model on the fly using a paper-based (LOM) machine. With the idea of translating prototypes into a conventional foundry process--or other conventional manufacturing methods--more interventions become possible. [This echoes to a degree the "prepared pianos" created by the composer John Cage in the 1940s. These interventions--bolts, nails, screws, etc lodged in the strings of otherwise conventionally designed instruments--led to a whole new musical form that tied the conventional "percussive" action of the keyboard to "aleatory" (chance) sounds.]
Art and Design Education Using "Digital Sculpture" Techniques
Just this year, a new course utilizing rapid prototyping has been developed for design students at Arizona State University. Professor Michael Nielsen from the College of Architecture and Environment Design, in collaboration with our PRISM lab (Partnership for Research in Stereo Modeling), has introduced a series of interative design problems that focus on consumer objects produced on rapid prototyping equipment. Students do 3D design work on computers located within the College of Architecture. Their finished digital models are sent to a networked Silicon Graphics computer in the Engineering Center that acts as the server for two RP machines--a JP 5 (a layered paper process) and a Genisys RP machine (a fully automated thermoplastic "3D printer").
There are opportunities for students to pass a number of continually modified designs through these machines. The design work will be done primarily in Form Z (which allows files to be generated in "STL"--stereolithography format).
Objectives for the class are as follows:
o Students will learn how to use "rapid prototyping" (RP) technologies (e.g., computerized milling, stereo-lithography, etc.) for "hard-copy" sculptures of complex three-dimensional forms.
o Students will learn techniques for merging diverse three-dimensional representationsbe they synthetically derived or sampled from existing three-dimensional sources.
o Students will learn to apply computerized iterative design principles in which the development of processes and products is seen as part of a cycle of refinementnot a finite, sequence.
o Students will learn how to integrate distributed computer-controlled operations via hi-end data transferparticularly in regards to cross-departmental networking and CAD/CAM design.
o Students will be challenged to work as part of a collaborative team and develop strategies for effectively integrating their design work with the work of others.
o Students will be challenged to assess qualitative issues with respect to design--in particular the relationship of hand-made objects versus those created with computerized equipment.
o Students will be encouraged to understand the significance of their activity historically and be prepared to discuss comparative strategies for prototype production in different technological and cultural contexts.
As an educator and artist, I see myself as helping to facilitate the class in a number of ways. From a broad perspective, I consider it as part of my role as PRISM's Co-Director to conceptualize how our research into 3D visualization and prototyping can serve the larger educational mission of the University. With respect to the class itself, I am one of a small team responsible for facilitating the actual production of the prototypes and helping to organize the lab. Beyond helping with the mechanics of the class, I am using the opportunity to better understand how technology can be integrated into the learning process. Key questions include: What are the benefits--and the liabilities--of having students tie their design processes to machines? Will these computerized techniques help or hinder their ability to work collaboratively? What impact will computers and RP devices have on the qualitative aspects of their work? Will the use of computers deepen their regard for alternative methods of working--or will it undermine our efforts to communicate an appreciation for all methods of creating effective design? Will the relative speed with which changes can be made to designs actually lead to better product? Will the production methods radically change the morphology of the objects produced? Can challenging new ideas from consumers be more effectively integrated into the design process?
Goals for Research and Teaching
What does the future hold? On a technical level, new machines are on the boards that will expand the capabilities of Rapid Prototyping technology into a broader range of scales and materials. Functional prototypes in materials ranging from hard steel to flexible rubbers are already possible. Can organic structures be far behind? Research on a theoretical level is currently being conducted into the next level of volumetric data storage. If a "pixel" (literally "picture element") represents the base level unit for typical 2D representations, the new unit of choice is the "voxel"--a neologism that combines the words "pixel" and "volume." Systems are envisioned that would be tied to enhanced voxel data sets that allow an operator to specify not only volumetric information, but material or physical property characteristics. Imagine, for example, a knife blade created in a rapid prototyping machine that would be able to be "custom crafted" to different specified levels of flexibility and density. More distant horizons hold out the possibility of "quantum computing" that...
In the area of Interdisciplinary Education, I am particularly excited about an integrated approach to techniques of visualization and prototyping that could serve as a vehicle for innovative teaching and a cross-disciplinary curriculum. These technologies lend themselves readily to hands-on approaches and "active learning." Further, as all design and output is part of an ever expanding digital record, the archiving of raw data and product will provide opportunities to pursue collaborative research across disciplines and with other institutions. Ideally, a critical discussion will ensue that looks closely at the formulation and interpretation of data and physical prototypes derived from that data. The theoretical (and legal) problems associated with digital imagery, virtual reality, and simulation, and the context-dependent factors that issue from particular methodological biases need to be explored by interdisciplinary thinkers not afraid to draw from historical precedent, the "science" of fiction, and studies in the humanities ranging from ethics, pscyhology, and education.
The demand for expertise in visualization, modeling, rapid-prototyping and manufacturing is exceptionally strong right now in both the arts and sciences. It is a growth sector for future generations of students. Access to equipment and training programs that extend these new technologies to those communities and constituencies who remain technologically impoverished are essential.
All of these new concepts and systems depend on the computer for its ability to translate quantifiable data into visual information. Using systems that are becoming increasingly accessible to the non-technologist by virtue of improved graphical interfaces and hardware, artists can now move with impunity (if not total freedom) in a domain heretofore dominated by computer scientists and engineers. While it would be a mistake to suggest that a lay individual can move easily into high end rapid prototyping, with ever improving GUI's and the ubiquity of learning tools for 3D digital media, it can be argued that "digital sculpture" has come of age. What about the rest of the culture?
Future individuals and communities may be distinguished by their ability to meet the challenge of form-making with methods that are quick and responsive--that meet the need for objects, abstract forms, or environments as they arise. New technologies that utilize transparent interfaces, reduce technological restraints, and encourage interative design processes and customized products have the potential for changing our 3D environment and making it more user responsive. Design and production by users-for users in real time, of course, depends on models of design, manufacturing, distribution, and consumerism that are only just beginning to take shape.
As with any new technology, the real challenge will be to not fetishize the machines or their products as ends in themselves, but to focus on the quality of the activities enabled by the technology. No one would dispute our status as consummate tool designers and makers. The challenge is for all of us to become better tool users.