Why Open Source Will Be the Catalyst for Next-Gen VFX Innovation
Simon Robinson, Cofounder and chief scientist at Foundry
January 22, 2020

Why Open Source Will Be the Catalyst for Next-Gen VFX Innovation

One of the most important trends across our industry is an accelerating move towards standards backed by open-source implementations.

In film and TV — which can rely upon multiple VFX houses working together for CG-heavy productions — collaboration can be hindered by disparities between the different facilities, particularly when it comes to data representations. Standardized representations make collaboration far more efficient, as well as reduce miscommunication and error.

Because of this, we expect open standards to continue to transform the VFX industry over the next decade.

Organizations such as the Academy Software Foundation (ASWF), of which Foundry is a partner, show that the industry takes the idea of a common platform seriously, while at the same time acknowledging that ‘openness’ is not a trivial exercise. It requires organization and real effort.

Studios today can still expend a large amount of effort on ‘reinventing the wheel’ and those that embrace open-source technology, can focus their own internal software development on things that make a difference to their creative output or efficiency. 

USD Will Be Key

An ever-growing family of open-source technologies has become available for studios, providing everything from unified color environments (OpenColorIO from Sony Pictures Imageworks) to image representation (OpenEXR from ILM).

 

For us, one of the most exciting of these is Universal Scene Description (USD). 

Developed by the R&D team at Pixar, USD enables the robust, scalable interchange and augmentation of arbitrary 3D scenes made up of different assets. This is an extraordinarily-hard concept to generalize, and has historically been something that each facility has needed to reinvent for itself.

The versatility and flexibility USD affords stands to be a game-changer.

USD allows multiple artists to work together simultaneously on the same scene or even the same asset, layering their changes together as they iterate on a project.

Furthermore, USD means that DCC applications from different vendors can share a common understanding of a scene. So, animation can be working on a scene at the same time as look development, even though the work is carried on in unrelated software and in different departments.

Real-time Rendering

USD is important for many reasons, but a further aspect is its potential to also play a pivotal role in realttime workflows.

Realtime is something of a buzzword in the industry at the moment, and encompasses a multitude of developments, which have the potential to drastically improve artist experience.

However, the fantastic recent innovations in realtime rendering technologies are hard to top. Access to instantaneous feedback can revolutionize the speed at which artists can iterate, and consequently it can change the way teams collaborate.

With USD as a standard, for the first time there’s a possibility of the industry adopting a scene format representation for current workflows that can also be an enabler for realtime workflows. 

A case in point is Hydra: a powerful rendering component of USD that also powers the viewport in Foundry’s Katana. USD describes the scene, and Hydra facilitates visualization of the scene, both in a universal way. So, you visualize it consistently and in realtime where needed; any software can use USD to take part.

Real-time Production

Looking at a broader picture beyond rendering, USD has other exciting implications for media production. 

The format could also be a vehicle that creates a bridge between story, previs, on-set work and post production.

Currently these can be fragmented processes, because the specific demands of each phase have led to the evolution of siloed software and data representations.

Having a world where data representation is continuous and uniform across the various stages of production, allows us to build technologies, which can be reused in multiple parts of the pipeline.

Let’s say you’ve come off set and you’ve done some virtual production. You’ve made a lot of live decisions, and your assets and editing tools may be perfect for that environment, but your effort so far may not translate easily into post production. It’s not unlikely that the post production teams will recreate your assets and scenes again.

But if a scene at each phase of production could be an iterated improvement on the stage before, then we could potentially maintain a continuity of decision-making through the development of a show. Any software tool typically used in one phase would be equally capable of being exploited in another. Perhaps then, the traditional linear stages don’t always have to be strictly sequential any more.

And it doesn’t stop there. We envision that open-sourced technologies, like USD, will bring major transformative changes to the future of VFX — exciting when you consider the undiscovered use cases and developments yet to come.

Foundry’s R&D team are already exploring the format for use in some of our most exciting research projects. All this means that, if you’re looking for where the next generation of VFX innovation will be born, open-sourced technologies like USD are a good bet.  



Simon Robinson is co-founder and chief scientist at Foundry (www.foundry.com).