Storing 3D VFX assets in an accessible way will reduce the cost and complexity of re-using those assets, make them easier to share between contributors in a workflow, and increase efficiencies across the ecosystem. An IBC Accelerator project is aiming to tackle this.
3D digital assets are an increasingly vital component in movies, TV shows and gaming - especially those heavy with Visual Effects (VFX) as well as digitally native entertainment like animation or video games.
Currently VFX and animation vendors create thousands or even millions of 3D assets on a project and the studio may need to coordinate between multiple vendors each handling parts of the project. While each vendor has its own internal asset management system and process, there are currently no industry-agreed processes or tools to normalize the storage of those assets either during production or when they are archived.
Storing 3D assets in a way that allows them to be searched and retrieved will reduce the cost and complexity of re-using those assets, make them easier to share between contributors in a workflow, and increase efficiencies across the ecosystem.
Accelerator Title: Usable VFX Archive
Champions: MovieLabs (Project Lead), Universal Studios, Paramount Studios, Sony Innovation Studios, Unity Technologies
Participants: Perfect Memory, Atmecs, Vidispine
This Accelerator Project was proposed by MovieLabs (Motion Picture Laboratories, Inc) as a proof of concept for a standardised ecosystem for the storage, sharing and archiving of 3D digital assets in the same way that media asset management systems already facilitate the handling of video and audio assets. In a similar way, a 3D VFX asset package would carry with it a descriptive metadata and a proxy for easy viewing of the package contents - to make it easier for a package of assets to be shared between vendors and their clients. The project will consider both operational and commercial challenges by minimising the inefficiencies inherent when multiple groups are working on and sharing the same assets.
MovieLabs is a non-profit technology research lab jointly run by Paramount Pictures, Sony Pictures Entertainment, Universal Studios, Walt Disney Pictures and Television and Warner Bros Entertainment to drive innovation and deliver tangible solutions to the challenges faced by Hollywood in transitioning to next-generation production technologies and distribution platforms. The organisation helps the industry develop next-generation consumer content experiences, reduce costs, enhance security and improve workflows. Universal, Paramount and Sony are supporting the Accelerator as co-champions along with Unity Technologies.
The project builds on a white paper published last year in which MovieLabs shared a vision of where the studios want to get to if they are to take full advantage of transitioning production to the cloud by 2030. While one of the ten principles in the paper suggested how archives should evolve, another identified that individual media elements need to be referenced, accessed, tracked and interrelated using a universal linking system. This Accelerator would help apply that principle to 3D digital VFX assets. An open, standardised ecosystem will also provide future-proofing by enabling various tools to understand, search and present/retrieve assets from an archive.
Project lead for the Accelerator is Raymond Drewry, principal scientist at MovieLabs. During his career he has been involved in several industry groups looking at related issues. These include both the Linked Content Coalition, a non-profit creating data standards for interoperable content identifiers and metadata; and EIDR – the Entertainment ID Registry – which provides global unique identifiers for singular and serial, commercial and non-commercial, audio-visual works.
Drewry sees the IBC Accelerator programme as providing a valuable mechanism for floating the idea and locating contributors with relevant insights: the project needs to combine VFX creation and asset management/archiving experience. Drewry is keen to ensure a good balance between the two disciplines.
The project will be managed in two parts.
In the first phase, participants will agree on a proposed set of metadata, formats and common definitions for 3D assets. This may include a review of the approaches currently being used by the different studios and, depending on availability, may also draw on a project that is being discussed at the Entertainment Technology Center (ETC) at the University of Southern California.
The project goes beyond the scope of EIDR - which only relates to identification and metadata – and will look more closely at the assets themselves which can comprise multiple sub-elements such as files for textures, rigging, animation and meshes and may also have related assets such as lower polygon complexity versions.
They will also look at how metadata is applied. Current practices are diverse and can involve embedding some metadata in filenames, some in embedded file metadata, and some in asset management software.
“We need some kind of standardisation for communication. And that’s really what this is all about – it’s standardising communication to make it easier to find items in a complex mesh of interrelated 3D components,” says Drewry. “The economic incentive is that if you can find something, you can share it, archive it and reuse it as a reference or component of a future project.”
The second part of the project will be building a simple, indexed repository that allows searching and retrieval. The goal will be a system that is independent of the actual storage mechanism so it is not tied to a particular cloud or local file system.
“Archiving is important because people want to reuse things,” says Drewry. “For a proper archive you need some standardisation on formats, and you need to call things by the agreed name.”
Reuse is increasingly important, not just within a single movie or episode, but for sequels, merchandising, promotional material and even for making plastic toys.
An accessible archive can also aid continuity between movies in a franchise. Drewry notes that fans pick-up on the tiniest differences or discrepancies between sequels or episodes.
When a character or artefact in a film becomes popular, easier access to related assets will open up new monetisation or promotional opportunities. At present, it is often necessary to recreate a digital asset from scratch because the original cannot be found or only some of the source assts are available.
With extensible metadata, archive searching becomes much more powerful: users can ask questions that are very hard to ask currently allowing them to surface existing assets for reuse or repurposing. Additionally, including still images or short videos of the assets in the archive opens up a wide range of machine learning-based search and discovery tools.
Drewry concludes, “While storing and sharing video files has become a fairly well organised process with multiple DAMs, agreed metadata and sharing systems, we’d like to enable a world where every VFX vendor can continue to manage their assets as they see fit but provide standardized tools for moving those assets in and out of their platforms.”
Although IBC2020 has sadly been cancelled, the IBC Accelerator programme continues to help broadcasters, studios and media & entertainment organisations collaborate to drive innovation throughout the year. By providing an environment for remote/virtual multi-company R&D, the Accelerators have come into their own in the current crises.
While there won’t be a physical presentation of the Proofs of Concept in Amsterdam, the Accelerators will feature in the IBC week of online sessions being developed for September. To find out more about the IBC2020 Accelerator Media Innovation Programme or to get involved, click here.