3d graphic design, 3d animation studios


Crash-landing near an iced over lake near Biosyn Genetics’ dinosaur preserve in Italy’s Dolomite Mountains, Owen Grady (Chris Pratt) and Kayla Watts (DeWanda Wise) are soon stalked by a Pyroraptor, which chases them on the ice.

This dramatic sequence in Colin Trevorrow’s Jurassic World: Dominion–which featured visual effects by Industrial Light & Magic–was made all the more complicated owing to the Pyroraptor’s feathers covering different parts of its body. In fact, to deal with the ‘Pyro’s’ feathers, and other feathered creatures in the film, ILM re-built its feather generation and grooming toolset.
Here, ILM senior visual effects supervisor and production visual effects supervisor on Dominion, David Vickery, and ILM visual effects supervisor, Dan Snape, reveal the brand new aspects of the VFX studio’s feather system, for befores & afters.

b&a: What I think is interesting about this character–the Pyroraptor–is that you had to deal with feathers, snow, you had to make it dive under the ice, and it had to be menacing. There were a lot of things going on.

David Vickery: Yeah, it was really tricky because of, as you say, the number of different atmospheric conditions and environments that we had to put it through. I mean, I think I’m okay to say feathers tend to strike fear into everybody’s hearts when it comes to visual effects, but for all the right reasons. We want to do stuff that’s really complex and really challenging.

But the complicated thing with feathers is finding a way to allow the artists to create and reiterate on work rather than spend their time just banging their head against technical challenges. We all want to make great images and create beautiful shots. If all you’re doing is trying to stop feathers interpenetrating or colliding or not grooming properly, that can suck all of the fun and the creative side out of the work.

We knew the Pyroraptor had to be out in the wind and the snow and the ice. We knew that it had to dive through the surface of the ice down under the water. It had to swim underwater and track with all those bubbles and the complexity of the underwater environment. And it had to burst through the surface of the ice. Actually, the script rather wonderfully had a line in it that said, ‘It lands on the ice, snow crystals forming on its feathers.’ And you’re like, ‘Yikes, what a cool image!’

b&a: Dan coming to you, when that comes to you and your team at ILM, where do you start with realizing those things, and feathers?

Dan Snape: We knew early on that there were going to be feathered dinosaurs. So we’d already started tackling the problem. Plus, the feather types of each of those dinosaurs were all slightly different, from, say, your protofeathers, which are more like the spines all the way through, to the Pyro, which we knew was full-feathered and had more like traditional-looking feathers.

That’s when we started looking at a new feather system, something that would be able to give us that flexibility, efficiency, but also creativity to change. It was then the script landed, and then speaking with David, we found out we were adding into that mix all of the different environmental challenges that had to go with it.

b&a: How did you start the process of building a new feather system?

Dan Snape: First we started looking at what had been our previous challenges with feathers. Like David mentioned, there were the issues with the technical side of things, of just being unable to iterate quickly. We wanted to find a more efficient way of doing things, so we built a feather system to do just that, allowing our groom artists to very quickly scatter and place guides across the body of the dinosaurs. They’re just working with guides, which effectively translate as the quills of the feathers themselves.

Once you’ve got that initial very quick scatter and coverage, that’s when then the artist, using a very intuitive paint system, will paint into those guides, attaching different styles of feathers to them.

Pyro has multiple different kinds of feathers across it. It has the head crest with very large, big feathers. It has the traditional wing feathers down on the arms, but then it has all those much softer downy, almost fluffy feathers, underneath the joints. With the new tool, our groom artists could quickly go in and assign the different feather types to each of the areas. The new system meant we could get very quickly to our very first, fully feathered, fully covered dinosaur with traditional feathers on there.

From that point, we then had to start discussing how those feathers would interact with each other. Built into our system was an auto-sim that would run a ‘detangle’ across the whole of the Pyro as well, which was just an amazing step forward. I mean, we were looking at a first version of the groom which looked great, and all covered, but there were all these intersections everywhere. Then the team said, ‘Just so you know, we haven’t run the auto detangle on this yet.’

David Vickery: A magic bullet.

Dan Snape: It was amazing. What that did was, it just took all of those guides, and all those feathers, and separated them. Then they settle, but they settle back to contact to give you the flat feather look. This means, at the groom stage, you are detangling, which then aids once we were getting into the creature team for sim’ing. There they’ve then already got a base sim that was a ‘detangled’ version.

At some point we were also discussing the silhouette of Pyro. We started saying, ‘Well, it needs to be sleeker in certain areas.’ And we were looking at references and birds, and what you sometimes get are areas of birds where they fluff up their feathers and they’re always a little bit bigger. Then in other parts of their body, it’s a much denser and slicker silhouette.

So, also built into the system was the ability to have a ‘flattening zone’ in certain places. What that would do is, it would take the curvature that had been applied to each of the feathers and flatten and push them. You could have, say, less gaps between any feathers. That was driven by masks. We could pick zones where we would flatten the area down. That allowed us to extremely quickly get to initially just a grayscale feathered version that we could then present to start getting feedback on the result very quickly.