AI Production and Software Tools
This session covers developments on improving different aspects of programme production. The presented papers are: the BBC’s work on a low cost virtual camera system for in-camera visual effects; on work at HFF and IIS, Germany, on exploring the use of AI assisted tools to support end-to-end visual effects production workflows that meet broadcast quality standards under real world constraints; and from WPP Production on a way to address the inability of current AI models to reason about the structural intent behind an edit sequence. There also are three strong supporting papers.
Presented Papers:
'OSMIA VP: A low cost field deployable virtual production system for in-camera VFX' BBC Research and Development, UK
'Exploration of AI-assisted asset tools in VFX production pipelines' - Hochschule fur Fernsehen und Film Munchen (HFF), Germany and Fraunhofer Institute for Integrated Circuits (IIS), Germany
'The causal edit: A pearlian framework for narrative logic in AI editing systems' - WPP Production, UK
Supporting Papers:
'Scaffolded AI: Mitigating the risks of AI-assisted development in multi-platform OTT' - Accedo, Sweden
'From script to breakdown: A run-based test harness for LLM-assisted script breakdown' Compos Mentis Productions, UK
'Beyond SAAS: AI assisted rapid application development in virtual production - three case studies' Fivefold, UK