The Best Choice for Simulating Plants
As simulations using computer generated graphics rise in popularity, various methods are evolving to represent actual conditions found in the real world. Presently, the methods being used to simulate organic materials, such as plants, are following three paths.
One method is known as "Raster" technology. This method is most familiar to imaging professionals, as photographs are scanned and turned into an array of little squares called "Pixels" (short for PICture ELementS). Each pixel represents a color, and depending on the level of the technology, any one image may be represented by thousands of rows and columns of pixels and millions of shades of color. This "pointillism" technique was first used by the french postimpressionist painters who would apply small dots of color to give the impression of a scene, to parallel how the human eye sees. The density and color range possible with this technology can far exceed that of the human eye, giving the illusion of a scene that is imperceptible from the original.
The second method is called "Vector" technology. This method is known to the CAD world, as it allows for accurate modeling of man-made objects, from small widgets to massive architectural structures. Based on the the concept of three-dimensional space (width, height, and depth - or X,Y,Z), this method follows the mathematical principals laid forth in the geometric relationships of points, lines, angles, surfaces and solids. It allows for exacting definition, and due to this exactness, is how a surveyor legally defines and parcels an otherwise random object such as land.
While raster is an outgrowth of the underlying technology, vector and the third method, known as "Fractals" aim at the mathematical represention of the structure of the object itself. Fractals are also a form of geometry. Benoit Mandelbrot, a research scientist working for I.B.M., coined the term Fractal as a way of describing repeating patterns that he observed occurring in many different structures. The patterns were nearly identical in form at any size and occurred naturally in all things. Whereas vector geometry excels in defining man-made objects, fractals are the most appropriate method for describing and defining organic materials, doing so by using very small finite algorithms and data.
Simulations vs. Presentations
The purpose for modeling plant materials generally follows one of two categories. These two categories could be broadly summarized as planning simulations and design presentations. Planning simulations are often technical in focus and scope. They might include growth simulations to determine height and diameter of the plant for interference checking, or seasonality to analyze aspect, habit and viewsheds. Presentation "renderings" take the abstractions of the planning process and communicate the designers vision for clients and others, often non-technical, to see. Presentations show the appearance, color and context of plant materials, bringing life-like realism to the rendered scene.
The three methods described, raster, vector and fractal, offer unique advantages and disadvantages with respect to planning simulations or design presentations. A review of each method will demonstrate where each may be effectively applied.
Growth Simulations
"From acorns, mighty oak trees grow" goes the old saying. While poetic, it can sure become a liability if the little sprout just happens to be under some power lines or in a narrow planting strip between a curb and sidewalk. This is where performing growth simulations can really save the day. While a good designer should have some knowledge of these issues, it may be difficult at times to anticipate all of the possible problems.
For analyzing space and area requirements, a designer modeling in 2D or 3D CAD is well advised to represent the height and diameter of plants as some basic form of vector geometric object, such as cones (for pine trees, for example), cylinders, spheres or some combination of these. This will greatly aid in checking for interference, and offers a very rudimentary simulation when performing a basic walkthrough, providing spatial context to the scene.
More complex simulations may be performed with fractal growth generators, which lets the designer see a plant's impact as an acorn or mighty oak or somewhere in between. This is useful to watch the rate at which the landscape will "fill-in". While allowing more complex analysis, they are often very difficult to set up, requiring specialized knowledge.
Raster images of plants may also be used to simulate growth, requiring the image be progressively scaled or "morphed". The limitation with this method is the fact that a juvenile plant often looks different than its mature counterpart. Overcoming this limitation requires the image be progressively altered and saved as new files by "cutting away" and scaling, to show increasingly younger versions. Once done, simulations that blend one file into another (called "morphing") yield surprisingly life-like results.
Seasonality
Plant materials have differing cycles with respect to spring, summer, fall and winter. Some stay relatively consistant, others completely lose their leaves. Plants on slopes will vary in habit, depending on whether the slope is facing north, south, east or west. Seasonality and habit simulations are especially important to regional and transportation planners. If a stand of trees is baren in the winter, the views from the roadway will be quite different than in the summer. Design is as much a part of revealing the non-obvious as well as showcasing the obvious. Plant materials, because of their varied natures, ideally suit this interplay.
Fractals are optimally suited to demonstrating this type of change, easily altering the image from one season to the next. Due to the inherent mathematical nature of the fractal technology, changes in habit may also be factored in. For example, a plant will take on a different habit in one type of soil versus another.
Raster images may also be used successfully here, but require that a variety of images be available to show not only the different seasons, but the various habits the plant may take on. This is important when a stand of trees is needed; you don't want each and every one of them to look the same. Of course, this problem is rather easily overcome by simply making alterations to an image, taking parts away and adding pieces, saving each new file as you go.
Vector representations are also somewhat capable of showing variation in seasonality, albeit very crudely. The other problem with vectors is performance. The more complex the model (often measured in "face count"), the longer it will take to render. Showing a pine as a simple cone shape is quite efficient. However, representing the same plant with all of its details would create a file so large as to probably crash the system.
Design Presentations
Design presentations are primarily intended for showing others what the design will look like in finished form. Here is where the appearance of plants as they will literally be seen is crucial. The colors and context of the actual plant materials play a vital role in the success of the overall design theme. Of key importance is the rendered scenes true-life realism. Where abstractions in the past were satisfactory for planning simulations and analysis, a rendering that is dull and lifeless will leave the viewer unsettled psychologically.
Because raster imagery reflects the process of human vision, it has the superior capability of conveying the elements observed in the real world. Using this method, the eye can be easily convinced it is seeing the real world. In fact, the best rendering technology today can make it virtually impossible to distinguish between whether a scene is a photograph of the real thing, or computer generated. To pull off this illusion requires the rendering be made from imagery taken from the real world; accurately processed so as to be suitable for computer rendering, thus allowing it to be seamlessly integrated into its new context.
Imagery taken from the real world also has the added quality of appearing life-like. This brings a familiar warmth and comfort to the scene. The mind uses association to make sense of the rendering, based on the memory of viewing similiar images out in the real world. When this essense of life is missing, the viewer is left with a cold and disassociated feeling.
Fractal imagery is very technically correct. It is based on the mathematical patterns inherent in the plant; portraying a structurally accurate representation of the actual plant species. Due to its close approximation of the real plant material, it is useful for design presentation renderings. Though the disadvantage with fractals is the difficulty of setting up a quality rendered image. It may take far too much time to "get it right" than is available for simply showing what it should look like, while using a raster image in its place is instantaneous. Inherent in its mathematical origins, the viewer readily knows that it is a "computer" plant, and not a "real" plant.
Vector technology is easily the least suitable for any design presentation. For rendering efficiency, the plant needs to look like a "lollypop". Even the most complex vector file renders out far behind the other two methods. It essentially comes down to what the designer's intent is with respect to the renderings impression. If you're a game designer creating an alien world, this may be entirely suitable. Yet, if on the other hand, your goal is to lend a warm and friendly feeling to the atrium of a new shopping mall or the foyer of a hotel for example, borrowing liberally from the real world will insure a well received and successful design presentation.
© 2009 Realworld Imagery Inc.
