Automatic Composition of Motion Capture Animation for Music Synchronization

Abstract

To enrich the music experience, automatic generation of a CG music visualizer is attracting more and more attention, where 3D animation is composed to synchronize with the music using motion capture data. In this paper, we present a novel approach for the above purpose, where both beat and intensity are employed to synchronize the motion with the music.We extend the conventional unstructured motion graphs to structured motion graphs (called weighted motion graphs) using motion beat and intensity, where a best path is searched by dynamic programming to obtain beat-level synchronization. Our objective function is designed for the following three aspects: motion quality, cost from beat synchronization, and cost from intensity synchronization. Our experiments with a user study demonstrate that the proposed approach can effectively generate attractive animations for music synchronization with much less computational cost than the state-of-the-art alternative.