Hi everyone! So many good questions to answer already! Thank you very much for all the compliments. We are obviously very proud of our movie J . There are a few questions of how did I get to this place in my career. That seems like a good place to start. Ill start by telling you something about how I got to my position and hopefully that will answer a few of the questions all at once. After attending San Diego State for Information Systems I quickly realized that the computer business wasnt for me (no-duh huh?), and I took my computer skills to art school where I studied graphic design. My first job in the business was at a studio that colorized black and white movies. Needless to say, this isnt what I thought Id be doing initially, but it was work using digital imagery on a computer.
Se7enthcin asks What was your first interview to get into the business like?
I was interviewing to be a colorist (painting frames one by one all day) and as it turns out the woman who hired me liked that I wore leather bracelets, REALLY. So, go figure, I wasnt hired for my talent, maybe she liked my portfolio as well. That was the opportunity that got me in the door and after two years of working with this technology we conformed it to be a digital ink and paint system for animation. That took me to London where I worked at Amblimation on American Tale 2 and Balto as a supervising compositor. I was in London for 4 years and then returned to Los Angeles where, after spending six months teaching myself 3D using the original 3D Studio (Max), I got a job at Rhythm and Hughes as a modeler. After working on Babe and Waterworld, I was hired by the then start up Dreamworks. And while at Dreamworks, I worked on Prince of Egypt, El Dorado, Spirit, and Sinbad as an effects artist and supervisor. It was after Sinbad that I joined Sony in the role of Visual Effects Supervisor for their first animated movie. So, mostly I was luckyeven my mom says this. But once I was in the door I excelled at what I did best, turning around images as quickly as possible, giving directors the most choices possible.
Porcupine asks did you know that you always wanted to do this? The answer is honestly no, I thought I would be in advertising or something. You have to understand that in 1988 there wasnt much of a digital imagery business so it didnt really seem like a likely career. But once I started, it was the only job for me. And everyday, I feel lucky to have this job.
Razorb asks What would you say is the most valuable thing you have learned on your journey that led you to Sony Imageworks? You will be pressured to decide which area of CG you prefer to work in, and you need to make this decision. The larger studios dont really hire generalists, rather they want to know if you are a modeler, lighter, rigger or whatever. Pick what you enjoy and excel at the most and then when you get into the company you can explore the other areas. Dont be afraid to change your mind once you experience a job that doesnt seem to fit you. Every discipline in CG animation has its benefits and rewards. Find the one that fits you the best.
From Buffichar: How long was the process between character design and the final 3D model before rigging? Roughly the answer is a month. We do al of our characters and all of our rigging concurrently, in other words as the first model finishes, it goes into rigging, and the next model follows subsequently. We may have five models in modeling at the same time with the previous 5 models in rigging. Obviously, the more important the character is to the story, the more time it gets in character design. We spent about 4 months on Boog and Elliot in modeling and most of our secondary characters went through in a month.
From se7enthcin What kind of render tricks did you have to use to cast shadows that were not geometry based? [color=LemonChiffon]This is actually a very simple technique called gobo matting. In traditional film or stage lighting, a gobo is a card with shapes on it that is placed in front of the light to give interesting shadows within the set. We did the same thing. We made shapes that would complement the design of the set and cast shadows from them and often turned off the real shadows that would have been cast from objects in the sccene. For example Elliots horns are complicated shapes and the shadow they cast is even more interesting. So when you see the movie, look for these shadows under Elliot as he is tied to the hood (bonnetsee I was in UK) of the truck. They arent there. [/color]
From Crossbones: Loved the film. I did get the feeling that as I was watching it was like still moments were incapsulated as paintings in my mind. The whole team at Sony did an amazing job!
Its great to hear that the movie left you remembering paintings because thats exactly what we tried to do. Without going into too much detail, we basically treated each shot compositionally as if it were a Photoshop file. We relied heavily on the composite to get our final approval. By that I mean we used Rendermans AOVs (multiple image outputs) to render as much as possible separately. All characters, lights, shadows, props, ambient occlusion, depth maps, and even regions within character were rendered separately and put together in the composite. This gave us the ability to tweak each element as much or as little as we desired. The lighting was added together in the composite resulting in an image that would be the same as rendering it all at once. It sounds like a lot of work but it gave us the ability to compose the lighting for each shot specifically.
In the pipeline at Sony were their any individuals that took shots on their own or ran together a series of shots and were responsible for those?
Our studio, and most CG feature studios, uses the CG Supervisor/Sequence Supervisor design that is, one GG Supervisor leads a group of artists to the completion of a sequence in a movie. The main benefit being those artists get a feel for that sequence and it gives the continuity and efficiency that we desire.
What kind of feed back to the animators get when they are animating (what kind of mesh and rig are they looking at)? Lets say something goes wrong after the fact in a fur simulation where the animator had intersecting limbs and it caused a problem, would the animator be able to fix the mistake later on in the production?
Lets start with the case of furred animals. Because the hair happens at render time, the animators dont see it when theyre animating. We give them what we call a volume stand in that roughly represents the volume the fur will fill in. This is just a guess so there will always be the potential for interpenetration problems as well as shapes the animator didnt expect. We use a system called Kick Back that basically, when its decided that a fix is needed, that shot goes on hold and the character gores back to animation to be fixed. This same example holds up for cloth and any unexpected deformations.
The fur system is fantastic.
How did you manage to render that amount of fur?
How long did a frame took to render, in general?
Imageworks has a really strong hair system that was developed over the course of the Stuart Little movies. Darren Lurie, one of the 4 CG Supervisors on this movie, and Chris Yee our Hair Lead, were integral into developing the look for Boog. The early tests that they provided were so successful that the appetite for what we could do with hair kept growing. It gave us a very high benchmark for what we should expect our characters to look like. We then had a whole team of artist who provided dynamics as well as shot specific combing for things like character interaction or even just design changes if a shot needed it. THEN
we made all the characters wet! That was essentially a whole other pass through hair look dev for all characters. This included new dynamic attributes and shaders.
Boog had the most fur and one HD frame of Boog took about 45 minutes. That wasnt that difficult to deal with. Our biggest problems occurred when we had to render many furred animals all at once. There is a shot of a bunch of squirrels up a tree. That took forever to render and often failed with memory problems. We kept trying to reduce hair count until you noticed they looked different and then split it up into Z comps when necessary.