3

Classic microfacet models for rendering account only for single-bounce light, leading to energy loss. We present a novel multiple-bounce Smith microfacet BRDF with a height-interdependent formulation, successive height preintegration, and an improved BRDF evaluation method. Our model can conserve energy well and achieves low variance with negligible bias.

We propose a framework which generates anime in-betweens while satisfying the technical standards of the traditional pipelines. This achieves state-of-the-art results, especially in complex out-of-plane motion scenarios. Our framework is unique and effective by filling the gap between existing video generation models and production-ready requirements.