Abstract
In this paper we propose an efficient new method for modeling the natural motion of cloth flapping based on a stochastic method for wind simulation and a P-type Fourier descriptor. Many have studied cloth behavior through modeling and animation, but only a few studies have combined a stochastic method for wind generation with the physical properties of cloth motion. However, the integration of artificial forces may alter the flapping and fluttering motion. We do not use physical simulation and forces; instead we apply the generated wind value directly to the cloth motion. We apply the 1/ fβ noise in the angles that define the curvature of fibers, using a Fourier descriptor because of its smoothing properties. This allows us to create a natural cloth shape. Since this approach does not require physical simulation based on equations describing the motion, we can easily tune the implementation using parameters that control the cloth motion in real time. Moreover, we propose a few simple techniques to improve the visual realism of fluttering, the appearance of swirling, and wrinkle effects.
