Magic Brush – A Novel Digital Painting Environment

Songhua Xu

2004-01-01

Abstract:A long cherished dream of both scientists and artists since the birth of digital computers is to be able to use the computer to produce beautiful art. With the ever increasing power of modern computers, the computer as a serious artistic tool is now within reach for many. One major research area in computer art is emulating traditional art media for interactive painting or calligraphy. A brush is many times more complex than a pen, and hence the problem presents a huge challenge for the interested computer scientists. We describe our solution to the problem in this article. Smith has written a good survey on the early painting systems [2]. More recently, several researchers have successfully implemented a virtual brush (or “e-brush”) that can mimic a real 3D brush for painting and calligraphy through physically-based modeling [6], [3], [8]. Compared with many existing 2D virtual brushes which often require the user to specify the contour’s control points and the texture of painted strokes (e.g., [11], [5], [1]), these 3D brushes are more natural to use, especially for non-computer specialists. A guiding principle for the design of an e-brush is that it must present itself as a familiar tool to the traditional human artist, because presenting a system that requires the designer to adapt, distract, or place attention outside of the actual task will hinder the creative process to which the designer can come up with ideas. With the e-brush we have implemented, the user will not be required to specify any control point or to adjust any parameter for the texture, but can operate the brush in more or less the same way as a physical brush. In a good virtual brush design, each step of the painting process should be simulated in high realism. Here simulation refers to both the production of strokes on the virtual canvass and the continuous visual display of the running brush. The latter is necessary if the user is to be able to feel the presence of a brush in order to have full control over it and to manipulate the brush as in real life. An even more perfect virtual brush should also provide a haptic interface, like what is done in [3]. In this article, we describe a new e-painting system [13], [12], [15], with focus on the modeling aspects—the geometry and the dynamics of the paintbrush. Our modeling of the paintbrush stands out among existing approaches because of its unique ability to capture the highly complex geometry of a physical brush as well as its dynamic behavior during painting with great accuracy. Figure 1 shows the overall architecture of our magic brush. Our modeling approach pays special attention to many Fig. 1. Architecture of our magic brush.