Simulation of clothing with folds and wrinkles

TLDR

Clothing conveys a character’s persona, and its draping, folding, and wrinkling must be accurately captured in computer‑generated simulations to preserve storytelling intent. The paper develops methods to match the behavior and appearance of real clothing on digital stand‑ins, producing realistic folds and wrinkles. The authors introduce a mixed explicit/implicit integration scheme, a physically accurate bending model with nonzero rest angles, an interface‑forecasting technique, a collision‑preserving post‑processing step, and a dynamic constraint mechanism to generate realistic cloth detail. These techniques collectively yield cloth simulations with abundant folds and wrinkles that significantly improve visual realism.

Abstract

Clothing is a fundamental part of a character's persona, a key storytelling tool used to convey an intended impression to the audience. Draping, folding, wrinkling, stretching, etc. all convey meaning, and thus each is carefully controlled when filming live actors. When making films with computer simulated cloth, these subtle but important elements must be captured. In this paper we present several methods essential to matching the behavior and look of clothing worn by digital stand-ins to their real world counterparts. Novel contributions include a mixed explicit/implicit time integration scheme, a physically correct bending model with (potentially) nonzero rest angles for pre-shaping wrinkles, an interface forecasting technique that promotes the development of detail in contact regions, a post-processing method for treating cloth-character collisions that preserves folds and wrinkles, and a dynamic constraint mechanism that helps to control large scale folding. The common goal of all these techniques is to produce a cloth simulation with many folds and wrinkles improving the realism.