Method Studios fills Le Mans grandstand for Ford vs Ferrari

Method Studios delivered hundreds of lively grandstands crowd shots for Ford vs Ferrari. Tim Riley, Head of Crowds in Vancouver, tell us about their workflow and how they pushed characters quality.

What was the size of your team and how long have you worked on this project?
We had a maximum team of 5 people, which fluctuated over the course of 6 months.
Golaem is so easy to learn, especially layout, we were able to borrow artists from other departments and train them quickly to help complete our crowd tasks.

Can you describe how many shots you did and which kind of shots?
We completed over 200 CG shots and supplied additional elements for 2D shots.
The main challenge was to add spectators into the grandstands at Le Mans 24 hour race. We had to place them around the racetrack, including in the pits to supplement the actors and extras on set, making sure to match various behaviours. While holding some props, such as umbrellas, cameras, etc.
At the end of the race, the crowd comes running onto the track, and surrounds the winning drivers.
There was also a sequence inside the Ford factory, which we had to digitally extend, adding CG factory workers.

How many characters are there in the shots?
20,000+ entities, made up from 24 different characters, with additional texture/shading variations.

In some shots the characters are very close to the camera, what did you do so that they can hold up even at this short distance?
Halfway through production, we were struggling to nail the lighting on some tough camera perspectives - showing the grandstand crowd at a low profile angle, where the sunlight clipped the tops of their heads. It just wasn’t working, with our polygonal hair geometry.
We utilized the new “Golaem fur” feature to add hair grooms to each character. We found this to be a huge improvement on the transmission of specular highlights, especially when backlit. It was great, not having to create anything extra for our rendering pipeline; we converted the grooms to Golaem fur, added to our characters, and simply re-rendered. We put hair grooms on ALL of the characters too, with hardly any memory overhead. 

Did you have to integrate FX with Golaem characters? What was your workflow?
There was a couple shots where we had rain running off some of the crowds umbrellas. We were able to export the umbrella tops as Alembic files, from Golaem, and passed them onto a Houdini artist to finalize. 

How did the Layout helped you making the shots faster?
It would have been very difficult to complete the project, in the time allowed, without Layout. 
Our workflow was to create caches as a grid of entities and then scatter these, using the Layout tools, on to our environment. This meant that any layout changes (which happened often), could be easily updated via destination terrain and population tools, without having to re-sim anything.

Anything else you would like to add?
I think the success of the Crowd Team was due to having a great toolset, from Golaem. We were able to integrate rigs, animation and mocap, from Maya very quickly. The layout tools especially, are very quick to create art-directable crowds, and easy for artists to learn. 
The viewport display, in Maya, is incredibly fast and high quality: it allowed us to get our crowd layouts approved, from a playblast, rather than a render. The render procedural also worked great, and installed quickly into our Katana/prman pipeline.

Learn more