Architects are in the business of designing spaces through the manipulation of real and virtual surfaces. During this design process, they need to continually evaluate the spaces they are creating among themselves and with their clients. They do so by using various techniques of representation of space. A partial list of these techniques include the perspective sketch, line drawings (plans, elevations, sections, isometric views), scale models, photo montage, computer animation and samples of actual building material.
Architects use different methods of representation at different stages of the design process, depending on what specific decisions need to be made. This is because representations are generally only good at telling one part of the story well. As the project evolves, decisions about different aspects of the design are made, at each point using a suited form of representation.
There are many phases in the development of an architectural project, from the "conceptual" to the "detail and finalizing" phase. At the beginning of the project, in the conceptual phase, the designer and the client discuss initial design ideas using rough sketches representing the basic geometric forms of the project. As the ideas mature, the project enters a new phase and gradually develops a vocabulary of forms. The conceptual geometric shapes become spaces made of surfaces and openings, and a first sense for the interior space takes shape.
At this juncture, the designer and the client need to make decisions about the way the spaces will "function" and "feel". More specifically, they need to evaluate (1) the sizes of the individual spaces, (2) the relative configuration of the spaces to each other and (3) the qualities and attributes of the individual spaces. These are spatial evaluations which, to be made perfectly accurate, would require that one be present in them. Ideally one would actually experience them as real places by walking through them. Clearly however, this solution is not realistic. It is inflexible, expensive and time consuming. As a result, architects have had to rely on representations of these spaces.
Historically, the representation technique around which these decisions have taken place is the scale model. Generally constructed entirely of one building material, such as cardboard, this scale model includes only the main architectural space elements, such as the walls, the openings, the floor, the ceiling and a removable roof (for better observation of the interior). If there are stairs in the project, they are often simply represented as ramps. The simplicity of the model and the lack of distinction between building materials helps decision makers focus on the main aspects of the space.
Scale models have been predominant tools for representing the feel of interior spaces for many years. However, there is a significant shortcoming in the use of scale models. They are small. As a result, to evaluate the "feel" and "function" of the space, users have to imagine or project themselves into the miniature model. While most of us are quite capable of imagining ourselves in these scale models, it is quite a different experience from actually being in them. How can it be certain users get an accurate sense for their own scale when they imagine themselves inside the model? And to make things worse, the problem is further compounded when the user has to imagine moving through the spaces.
Fortunately, advances in technology have brought considerable improvements to the problem of scale and movement in scale models. The first such improvement was the introduction of miniature cameras into the scale models. Typically, these cameras move about at eye height in the model, and their image is transmitted to a television monitor. More sophisticated versions have an apparatus which allows participants themselves to control the movement of the camera around and about the model (Bosselman 1987, Sasanoff 1967). With this new technique, the task of understanding the spaces is greatly simplified because it removes the need to imagine oneself in small scale. As a result, evaluations about the modeled spaces have become more reliable.
While the use of mini-cameras was being developed, parallel advancements in the field of computer graphics and animation began replacing the role of the physical scale model. Computer models are more flexible because they can be modified at very little expense. For many years, this technology suffered from limited graphic rendering and computer speed. Early computer models consisted of line drawings, and the "hidden line" was not hidden. Faster computers made it possible to display solid shading of polygons, and still faster and better programs added advanced rendering techniques, such as ray tracing, transparency and shadows. While these sophisticated renderings take a lot of time to compute, when transferred onto video tape, they offer convincing walkthroughs (batch computer animations).
As the power of computers increased, the rendering time decreased. Soon, given a simple model, it became possible for a viewer to change their viewpoint in a computer model in near real time, that is , 10 to 20 times per second. One such program is Virtus Walkthrough(TM). By sacrificing high rendering detail for speed, it allows viewers to move their viewpoint throughout the model in real time on the Macintosh platform. More sophisticated platforms, such as the Reality Engine, which runs on the Silicon Graphics workstation, are fast enough to render very high realism scenes in real time.
There is no doubt that both the advancements of technology in the mechanical field (mini-camera) and the computer field (real-time walkthrough) have improved the task of representing interior spaces. They allow viewers to better understand how the modeled spaces would feel because they have placed the position of the viewpoint inside the model, they allow the viewer to choose how they want to see the spaces, and they facilitate viewers in understanding how it would feel to move through the space.
However, this viewing perspective has created a new problem which did not exist before, and it has displaced a second. The problem it has created is that it has reduced the 3 dimensional spatial information of the scale model onto a 2 dimensional medium. In effect, the specificity of spatial information which was available in the scale model, now has to be interpreted from a 2 D representation. It has also displaced the problem of scale, rather than resolve it. While we take much of this for granted, we perform an important transformation of scale when we look at television monitors. We understand that the actual size of objects seen on the monitor have no relationship to the size of the monitor itself. The view of a building which only fills half of the screen is understood to be many tens of feet tall.
Furthermore, the participant is only remotely coupled to the model because they control their viewpoint using a kind of joystick device. They are more coupled to real settings and to a certain extent to scale models because they change their viewpoints by moving their head and their body.
For the reasons listed above, monitor-based representations are far from perfect for describing the "feel" of proposed architectural spaces. To make them closer to perfect, they would have to offer a more convincing illusion of depth, a more appropriate sense of scale and a greater coupling between the process by which people explore spaces in simulated and real environments.
There have been tremendous advancements in the field of computer rendering in the last several years and a merging of technologies which together, could solve all of the problems related to existing forms of computer based simulations. This new technology is called "virtual reality" for some, "cyberspace" or "virtual environments" for still others.
Virtual environments create a powerful sense of immersion within a computer model. Participants are immersed and surrounded by information which is to scale and which is 3 dimensional. The interface is very intuitive to use for exploring virtual environments because it is tightly coupled to the way people explore real environments. Viewers can look around in the model by turning and moving their heads, as they do naturally in real spaces. Because of the specific attributes of virtual interfaces, people develop a sense of actually being somehow present inside the model. And with this sense of presence, viewers could potentially, for the first time, perceive the modeled spaces as they would the real spaces. Virtual environments are poised to be the perfect representation tool for helping architects make decisions about architectural spaces before they are built.