Kiel Figgins - Character Animator
Film
Broadcast
Personal
Resources
Reel
Resume
Contact
Store
Butterfly Swarm - RnD / Previz / Animation
In short, this project was hard. The original idea of having 'about 10 or 12' butterflies cover a shoe and wrap around it was straight forward enough. I had done something in the past that was similar ( http://3dfiggins.com/Broadcast/ShoeRig/ ) which is why the client reached out to me. As I started doing the initial tests and RnD, the butterflies got a bit smaller, then they needed some bg ones flying around, then some already on the rock, then..., then... it wasn't long before the number had spiralled to over 300. But by this point in the project, there was no turning back, we couldn't bring on more people, and I was in a hard spot for money, so deeper it went.
Getting the actual tech working was straight forward. The control effector looked for a node in the proxy rigs for each butterfly and ran specific color changes in the wing shader and triggered specific flap or fly animations. The proxy rigs in my scene had the same names/pivots as the controls on the anim rig, so a transfer script was used to get the render rigs to line up to the previz I was doing. And here was one of the first hard lessons we were about to learn, the anim rig and shaders worked fine for the color blends and line up... when it was tested on 10 butterflies. As the number increased exponentially, a render bottleneck hit hard and only got worse as the shaders got more complex, the rigs got heavier and the poly counts kept raising.
That was only the first of many hits during production. My machine, though decent at the time was quickly being unable to support the scene, let alone all the anim rigs, so proxies needed to be created to allow faster playback. As the density of the butterflies increased, changes became slower. The idea changed multiple times about if the shoe was already there, was part of the rock, formed later and so on. Colors where added, audio timings changed, the spot got longer and longer, and so many others made progress slow.
Near the end, I was working unbearable hours and that caused rifts in my family that lasted long after the project ended and the money was spent. Looking back, there are countless things that would I would do differently. Testing at more regular intervals, clearer art direction before signing on, and such all make sense, but it's hard to stress how much money was a factor at that time in my life and how straight forward the project seemed, especially since I had done similar work previously.
Though I can't change how this project went, I was able to take those hard learned lessons and apply the tech and approaches to later projects, such as for the Pixels title card previz ( http://vimeo.com/382815952 ), for swarms of ships in Gaurdians of the Galaxy V.2 ( https://youtu.be/2cv2ueYnKjg?t=49 ) while at Framestore London, and countless other bits pulled and re purposed along the way.