Jul 2026· Proceedings of the Digital Production Symposium· pp. 1-8· 0 citations· 3 references
TL;DR
A pipeline in which Unreal Engine 5.6 is the final rendering and compositing environment and a dual-format animation pipeline combining FBX and Alembic with per-asset tagging and a data-driven Blueprint architecture is built.
Abstract
Dwarf Animation Studio is producing Twisted, an 80-minute animated feature directed by Lino DiSalvo, on a 16-month timeline including pre-production. To meet this schedule without compromising creative ambition, we built a pipeline in which Unreal Engine 5.6 is the final rendering and compositing environment. On our earlier production using Renderman, a single frame took 35 to 60 minutes to render and needed over 200 dedicated render nodes. On Twisted, a fully post-processed 2K frame renders in approximately 3 seconds (Figure 1) on fewer than 10 farm nodes. When render latency becomes negligible, the organizational logic built around managing that latency breaks down. This paper presents the technical systems we built to make this work: a multi-shot lighting workflow on nested Unreal Level Sequences with automatic shot group generation; a dual-format animation pipeline combining FBX and Alembic with per-asset tagging and a data-driven Blueprint architecture; a modified USD exporter that outputs environment data once per sequence with lightweight per-shot override files; a painterly look development pipeline built on custom shaders; and a performance strategy within a hard 16 GB VRAM limit on an RTX 5080. It also presents, with equal candor, what broke: the collision between departments that simultaneous data access made inevitable, and the organizational friction of migrating artists from a centralized network-drive pipeline to a version-controlled local workflow.
This work presents LiveAnimate, to their knowledge the first animation system to combine real-time streaming with stable long-form generation at billion scale, built on a 14B-parameter video Diffusion Transformer (DiT).
Yuxuan Zhang, H. Xiong, Yubo Huang et al.· 1 citation
Wan-Animate-2 is presented, an end-to-end character animation framework that directly consumes the driving video within a redesigned Diffusion Transformer and achieves superior motion fidelity and identity preservation by eliminating intermediate motion extractors entirely.
Guangyuan Wang, Liucheng Hu, Dechao Meng et al.· 3 citations
Executable programs offer explicit control over how images and videos are constructed, but generating runnable code is only the beginning of visual creation. A program can execute correctly while violating the requested composition, appearance, or motion. We define this discrepancy as the Program-to-Visual (P2V) gap an...
Hao-Yu Zhao, Zi-Hao Zhang, Xu-Dong Wang et al.· 0 citations
Magpie provides a system-level implementation path for applying generative models to real-time game rendering, and preserves gameplay designability and reproducibility, and reduces the dependence of early game prototypes on complete visual assets.
Xiao-Yu Zhan, Xin-Yu Wang, Xiao-Hong Zhang et al.· 2 citations
In modern digital animation production, with continuous technological advancements, especially innovations in computer graphics and artificial intelligence, the efficiency and quality of the animation production process have become key factors in the competitiveness of production companies. However, despite the rapid d...
Bo-Long Jiao· Journal of Engineering, Proj...· 0 citations
StateFlow is presented, a state-centric framework for generative previsualization that uses an editable 3D world to organize scene structure, evolution, and cameras, while off-the-shelf video models enhance visual quality when higher fidelity is desired.
Yuyang Yin, Zixiang Li, Longxiulin Deng et al.· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.