FMX 2013 - Le Big Shift in VFX
Date: 23.Apr 2013 | Reading Time: 2 minutes, 3 seconds
Topics discussed center around what the industry can do to improve interoperability and workflow to strengthen the business instead of running it into the ground.
Open Data Platforms
Rob Bredrow took the lead by talking about the work he and SPI has worked on to create a good open standard onto which companies can build to achieve something greater. Alembic, OpenColorIO, OpenEXR to name a few.
Before every company needed to reinvent the wheel in-house to set itself apart. Now they can work on a common standard which helps with interoperability between companies, which is something that is required in today’s industry.
We have moved from “secret sauce” to common baseline, which also includes the game industry. A convergence of VFX and games in this respect seems inevitable to Cevat from CryTek.
The words “vector of cross-pollination for these industries” were uttered. That should tell you pretty much everything you need to know.[^1]
Cloud Based Solutions
Cloud based computing power is something that is a topic everybody was interested in, even the big houses, who usually have several thousand render farm computers of their own. There are immense draws to this kind of workflow. From lower overheads, due to savings on machine, administration and licensing costs to being able to ramp up render power quickly when you need it during a deadline crunch.
This interest is something that is shared equally between small studios like ours and the big boys. And only expected to expand in the coming years. Ludwig von Reiche talked a bit about cloud computing applications in development, running on Amazon Elastic Computing among others, which he expects to come out within the next year.
While there are already solutions that offer Amazon Cloud rendering, they are usually cobbled together requiring a small science degree to figure out. We are talking about more accessible solutions.
Rob Bredrow argued that there are really two sides to cloud computing. One is the reduced cost due to less inventory and running costs. The other side is to add a lot of render power on demand, say 1000 or more machines to render a shot in an afternoon. That might cost more, but speeds up the creative cycle and might be worth it from that perspective.
: That’s around where I got bored and lost a bit track of the conversation.
Subtitle: Panel Discussion with Marc Petit (Autodesk), Rob Bredow (SPI), David Morin (Autodesk), Don Parker (Shotgun), Ludwig von Reiche (NVIDIA ARC), Cevat Yerli (Crytek)