Cycles. No, not the kind of cycles that you get by plugging in nodes that drive the nodes that drive the nodes doing the driving… I’m talking about the cycles that I see from year to year, project to project, from leadership team to leadership team. Just like how the industry goes through cycles as a whole (from hiring frenzies to layoffs to more hiring frenzies), those cycles can also be found within the details of the work done throughout the industry, and for the focus of this conversation, I’m talking about the changes of rigs over the course of multiple animated productions.
As a rigging supervisor, part of my job is to always evaluate what we’ve done; what has worked, and what has not. What animation needs vs what they want vs what we can actually give them. It’s always a balance – that never changes. Budgets slim, schedules get tighter, and workflow needs to become more efficient. As riggers, we adapt and move on. Stepping back, however, and looking at our changes over the years, it’s easy to see our own cycles of repeats and back and forth’s between old and new.
When we reevaluated our workflow and modular rigger system after Open Season 2 and the Kung Fu Panda commercial work, the biggest focus was speed. We managed to simplify the rig internal complexity while keeping the functionality the same, in order to gain a large rig speed increase.
When we moved to the Madagascar commercial work and the Escape from Planet Earth rigs, our focus went to flexibility and workflow. We had a huge increase of workflow tools that cut our rigging schedule down, so a shorter schedule was much more manageable. We also had a big epiphany in new approaches to rigging. Our modules changed a lot and we came up with different approaches to rigs. Part of this was the nature of the beast – the Escape job had a lot of abnormal characters to rig, which caused us to come up with a lot of different module types.
With Open Season 3 (another rig speed increase, mostly an attempt to de-plugin-ize the rigs), we continued the cycle. The Looney Tunes rigs were crazy. Complexity increased (nature of the project and what the characters needed to do) and by the 2nd set of Looney shorts (Elmer and Daffy), the rig speed took a hit because of the added complexity. They needed more controls, more ways to shape the characters in non-standard ways. Our attempt at taking the rigs to the next level was full of experimental ways of rigging and wasn’t focused on the most efficient ways at the time. Deformers stacked on deformers. Using soft mods pre- and post- deformation as controls. Lattices on lattices. It had cool results, but at a hit of slow rigs.
And so the cycle continues. From the Ice Age Xmas special to the Despicable Me Universal Studio ride and multiple projects in-between. Make it faster. Simplify the controls but keep the same amount of control. Now increase the controls, increase the geometry resolution for more detail… and then again, make it faster. Adapt. Free Birds rigs went in one direction, while Book of Life rigs went in another. Winged characters vs wooden puppets that articulate. A cycle of trying to simplify and consolidate controls ran into a cycle of adding complexity (on a tight schedule + small team too!).
It’s a small industry, so it’s easy to have conversations about X studio doing this and Y studio doing that. You get to hear lots of opinions from animators all over, and filtering that info down into “the ideal way would be THIS” is key. One studio like attributes-galore, while another wants pickable controls and limited attribute lists. What way is correct? Neither. And both of them. It’s up to your animation supervisor and the animators… communication is important! Adapt and improve!
This is not true. Right? 🙂
And so, as we approach starting on our next animated feature film, it’s time for a reset. Where are we in the cycle? It’s time for faster rigs. More innovative approaches. Ribbons are out while bezier math is in. Maya 2016 parallel computing is providing a great speed boost right out of the box. Less joints in a rig does not equal simplified and better control. Sometimes you need more joints, but better ways to control them all with less controls.
Having a super flexible modular system has been extremely helpful through all of the cycles above, and so with this cycle, it’s been time to do a reset. Revamp the code and backend. Push the old mel code aside to replace with python code. Make the modular rigger smarter, more adaptable, more consistent between characters and projects, and less prone to user error. Make workflow easier – automate joint placement where you can, skinning where you can, and provide tools and plugins to help when automation is out of reach.
And most importantly, have close interaction with your animation team to discover how you can help make their lives easier as their seconds per week quota climbs and everyone’s strive for better quality climbs as well. It’s a team effort!
So what’s the point of pointing out the cycles? Not every project will be perfect. You’ll do things that you should probably do differently the next time around. Multiple times. People’s opinions and desires will change, and you will need to adapt. You will need to think outside the box, but don’t forget to look back inside the box every so often… you might find some good ideas that are still valid. This is all part of being a rigger, character TD, articulation artist, and wizard. I’m looking forward to being able to have our team share what new things we’re creating with the community, hopefully in the not-too-distant future. Until then, happy rigging 🙂
– Josh Carey