↪︎Animating the Scramblers the normal way wouldn’t work, that much was clear from the start - creating the movement of the tentacles was difficult enough, and in scenes featuring dozens of the creatures it became simply impossible.
↪︎For the first test we tried creating tentacles that simply followed the Scramblers’ bodies. We wouldn’t get far with this set up, but at least it solved part of the problem in group scenes.
We just wanted to share the progress we’ve made with the illustrations and, for the first time ever—the Scramblers in motion.
↪︎WOW. The look, the movement of the scramblers is great. That aesthetic is exactly what I imagined. All I could say is—and this is really minor—all the scramblers you've rendered are juveniles. As they get older, the central disk thickens and becomes more spheroid; the arms migrate off the equator and end up distributed all over the body. So the adults don't look as much like earthly brittle stars as these guys do.
But man, I can't get over how gorgeous the rendering and the movements are. If you have any more of this stuff, I'd love to see it.
↪︎But what about the shots in which the Scramblers interact with their surroundings, rather than just wave their tentacles about? After reviewing articles on Octopus locomotion (our closest reference), and several weeks worth of coding, we were able to create a behavior that seemed convincing enough.
↪︎The movement of the body was still determined manually. Then, based off this movement and our algorithm each of the tentacles independently searches for the most convenient surface area to attach too, as well as when to detach and search for new footholds.
The system that controlled the “normal” Scramblers and its upgrade created for Scramblers interacting with their environment.
↪︎Time to send off another preview, featuring the “playground” in which we tested the Scramblers movements.
↪︎Not a big update but I thought you might like this one. Some work in progress on the system that controls the Scramblers leg placement and animation.
I realized that I have dozens of these guys in one of the shots so animating them manually would take weeks. 85 lines of code and a few pain-in-the-neck hours later each Scrambler can now decide where to place each leg on its own. Not exactly superhuman intelligence but still much better than what I could’ve done by hand.
↪︎This is really cool (I especially like the little vector/node thingies). But I've already seen a whole wall of those guys squirming and writhing in the trailers, and they looked incredible months ago without any scrambler-level AI. Why is this necessary?
Sometimes we could conceive of things and still not see them, although they stood right before us.