Insights

Technology at Image Engine: A Q&A With Andrew Kaufman

Software Engineering Supervisor Andrew Kaufman describes the evolution of Image Engine’s pipeline, from early investments in open source technologies to workflow innovations that drive the studio today. 

June 14, 2021

Image Engine is known for having a powerful visual effects production pipeline. How was that reputation established?

Without question, our reputation for technology comes from the consistency and quality of our digital content, and the credit goes first and foremost to the skill and passion of our artists and supervisors. We are rare in that we can produce this level of work while maintaining a modest mid-sized studio culture, and this more than anything speaks to the power of our pipeline.

The first building block was a dedication and investment in technology as a core foundation of the business. This investment allowed Image Engine to assemble a team of industry veterans and newcomers with new ideas, and to establish itself at the forefront of Canadian VFX, at a time when common thought was that only large corporate studios can create award worthy content.

We’ve followed on from these initial successes, delivering quality content year after year, while regularly presenting our methodology and technology at worldwide events like Siggraph, FMX, and DigiPro, local events like Spark and various DCC User Groups, as well as university outreach to technology and art students within British Columbia and abroad.

What motivated Image Engine to prioritize strengthening its technology? What did the studio observe in the industry and how did that lead to this focus?

In the early 2000s, there was an implicit acknowledgment that most VFX studios use proprietary versions of what are roughly the same technologies. At the same time, several DCCs had come to market to meet a significant subset of those needs. It seemed clear that we could build a functional framework for VFX based on a mix of 3rd party and open source software, so long as we were willing to invest directly in expanding the open source ecosystem ourselves.

What followed is 15 years of active investment in open source VFX software engineering, coupled with development of proprietary tools and workflows that build on top of our open source foundations, enabling us to punch above our weight in the global VFX market while maintaining the culture that draws great artists and technologists to Image Engine.

How do our specific open source and proprietary technologies help artists and the studio? What do they make possible?

Cortex  and Gaffer  are the two major open source technologies that Image Engine is directly involved with and they are both foundational components of our pipeline.

Cortex has certainly evolved over the years, but its primary focus remains the same; it provides the data structures and algorithms required to produce application independent VFX content. Cortex enables us to retain a unified approach to VFX while adapting to the ever growing technology ecosystem. Over time, we’ve incorporated most industry leading open source libraries into Cortex, including Imath, EXR, OCIO, OIIO, VDB, Alembic, and USD, as well as live data manipulation in the top DCCs, like Maya, Houdini, and Nuke.

Gaffer is a first class DCC in its own right. It puts large scale 3D & 2D proceduralism into the hands of artists and TDs. It is built on top of Cortex and the other associated open source libraries for VFX, leveraging the data representations and algorithms in those projects, and wrapping them into a multi-threaded computation framework and procedural node graph interface.

We utilize these two open source foundations to build other proprietary technologies, such as our  robust asset management system, called Jabuka, which provides consistency in data flow across all departments and all productions in the studio. Jabuka enables artists and TDs to evolve workflows from show to show, build reusable templates, automate repetitive tasks, and easily propagate changes across hundreds of shots.

What are the current realities of the visual effects industry today that are driving workflow innovation at Image Engine?

The emergence of deep learning, realtime technologies, and cloud scalability have pushed automation to the forefront of the industry. We’re seeing both client and internal expectations grow exponentially as new software campaigns roll out; our technology stack has to adapt faster than ever to meet this demand.

With so much 3rd party and open source software available in the industry today, we’re constantly re-evaluating past decisions, deprecating existing tools and workflows, and strategically reinventing ourselves from show to show. Simply put, we’re sticking to the common framework and ethos that has defined Image Engine for decades.

To accelerate this exploration we’ve created a dedicated Workflow department, charged with discovering and advocating for new approaches to common studio challenges, as well as expanding our capabilities to meet the ever growing demand for creative digital content. By coupling Workflow experts directly with Software Engineers we will be able to explore ideas quickly and thoroughly, and pivot our approach to technology as the ecosystem evolves.