Meet the Artist: Spencer Cook

Hi Spencer,
am so happy that we have the chance to chat a bit with you ... thanks to cgsociety for that ^_^

I didn't prepare any questions yet ... but till I post mine I wanna say hello and pass a little 2 questions

- How long did the shot take from you to animate ? I know that there was some complex scenes and some are more simple ... but am focus on the complex side ...
- Did every animator take a shot to animate it by himself ? or there was more than one to animate the same shot ?

thanks again for you kind reading ...

- Yusuf

Hi Spencer,

First of all thanks for responsing our questions, so here's mine:
How long did it take to create the character rigs, especially the one of Spidey?

And as CB_3D mentioned, it would be nice to see a lowres model of Spidey

Best regards,
okazaky

Hey Bonedaddy,

Thanks for the questions. We called this creature "the goo" when it's crawling around by itself. It becomes Venom once it merges with Eddie Brock. To answer your questions;
1.Yes, the process was mostly hand keyframed. We tried doing it procedurally in the beginning but we found out quickly that it was going to be tough to get the organic quality that Sam was looking for solely with that technique. The basic rig was mostly spline IK for the tenticles (we called them "pods"). Koji Morihiro and Ryan Laney set up a system of maya shelf buttons that we could use to import pods and build the rig on a per shot basis. The pods included attributes for bulging and pulsing that could also be keyframed. After the keyframe performance was done the fx team added more layers of procedural goo to give it more organic detail.
2. You're pretty close to the mark with your comments here. The goo was a fairly specialized task within the character team. The animators that really excelled at it were the ones with strong technical skills as well as great character animation skills. Koji really helped set the standard for this character. He's a great rigger as well as a great animator.
3. The Spider-man rig has changed very little since the first movie. We have mostly improved the selectable controls to make it more animator friendly but the inner workings of the rig are essentially the same. Our riggers on the first movie (Koji Morihiro, Aaron Campbell, JJ Blumenkranz and Albert Hastings) did such a great job that we haven't found it necessary to make any major changes to the base rig and weighting.
4. I don't know that things will change too dramatically in terms of who does what. In a large studio like Imageworks everyone is very specialized. Partly because everything moves forward in parallel (ie; animation test are happening at the same time as modeling and rigging) so there aren't a lot of artist that could, or would want to switch tasks. Sorry, I don't know if that really answers your question.

5. Time is always the enemy in this kind of work. Sometimes you don't have enough and sometimes you have too much(some shots can be over-worked to the point where you've passed the sweet spot and are not really improving it)



thanks,
Spencer

Hi Azurelle,
Thanks for the kind comments and good questions. The ideal, of course, is a movie with an interesting story and characters we can care about, otherwise we get a feature length demo reel. I hope that audiences today are more interested in good stories and exciting action scenes(whether stunts or fx) rather than looking for the next great fx sequence. I always get a little nervous when reviewers talk too much about fx for a particular movie. General audiences should be wrapped up in the story and not be on the look out for the newest fx.

thanks,
Spencer

Hi BlackStorm,

We use Maya for all of our character animation. Other software used on the show included Houdini, Renderman, Katana, Bouju, Shake, Syflex and a ton of proprietary tools and software developed in house.

thanks for the question,
Spencer

What is all this non-sense you guys are talking about? Renders, rigs, motion capture! scrap that! The reason the movie is so realistic is that Mr. Cook and his associates actually went through the trouble of setting up a multidimensional space-time machine and went to a dimension where Spidey and company actually exists, the so called Einstein-Rosen Marvel dimension. There, they were able to convince Spidey and his arch enemies to provide some action for a few million Hollywood bucks plus hot chicks, which were then shot (the action, not the chicks!) from a complex set of cameras strategically positioned. Really simple.

kidding aside, congrats for the amazing work!

Hi GreenArrow,
Thanks for the kind words, we always try to learn from our mistakes and improve with each project. It's great to hear that you appreciate our efforts.
To answer your question about "plates". You are correct. A 3D replica of the environment is made but only enough to define the basic geographical parameters of the real environment. The purpose of this is to recreate the camera move for that shot in a 3D scene file, the process is called a matchmove. The 3D environment is used by the matchmovers to judge the size and distance of objects in the real environment. When these variables are known they can accurately match the virtual camera to the real camera allowing our animated character to fit into the plate realistically.
As to your other questions;
1. Every shot is different. Some shots can be done in a week or less. Other shots, like
the first shot in Sandman's birth sequence (over 2700 frames long!), took many, many months to animate. We budget every shot at the beginning of a show, which is basically an estimate of how long we think it will take to animate, add fx, light, comp, etc. Sometimes we come in under that estimate and sometimes we go over.
2. We always try to work as efficiently as possible. Addressing the director's comments and showing an updated version of the shot as soon as possible is usually a top priority. Any technical or administrative issues that get in the way of that need to be dealt with quickly.

thanks,
Spencer

Hi LATROMMI-SUINEG,
Thank you, it's always great to hear these comments. I'm very proud to have been able to contribute to Sam Raimi's wonderful films.

Spencer

Hi akhenaten,
Thanks for the compliments. The effect you are referring to is from a sequence in Matrix Reloaded that I didn't work on. All of the bullet time effects were done at ESC. At Imageworks we mostly handled the sentinels and ships in the tunnels. However, I'm pretty sure those were digital models in that particular shot. If anyone knows differently please correct me.

thanks,
Spencer

Hi Spencer,

My question is a little off beat. I was wondering how important was 2D animation on your journey to your current success. I'm asking because I started out in 3D but have since been drawn to traditional animation. Is there a real benefit these days to have a stong 2D background and how realistic is it's foundation in 3D these days.

Thanks.

Hi JamSession,
Good question. I would not want to lose the IK/FK switch in our rigs. When I'm animating, I prefer posing a character using FK controls as much as possible. The exception, of course, is walking, running or crawling, IK is essential for actions like that. Generally though I feel like I can get more predictable results from FK. IK joints sometimes solve in an unnatural way that needs to be corrected. In our rigs, the animator can choose which one to use at the start of a shot or animate a switch during a shot ie; Spidey is swinging (FK) then lands on the side of a building (IK). I know many other studios have this option in their rigs, Imageworks didn't invent this.
In answer to your second question, we didn't use any mocap for our characters except for facial performances. The reason for this is that usually a shot is animated because a person couldn't really do it. We tried using mocap of a gymnast on Spider-man 1 but it looked like what it was; a trained gymnast going through a routine. Spidey's moves and poses are very specific and atlering mocap info was very time consuming. One thing we do use a lot is reference footage. Sometimes it's video of acrobats and gymnasts or nature footage of animals running, jumping and crawling( I looked at a lot of video of lizards and frogs crawling for ideas when I animated the wall crawl shot in the first movie). We also videotape the actors whever possible and sometimes photograph ourselves acting out a scene. Sandman's birth sequence could have been done with mocap but Sam's ideas for the performance changed over time and it wasn't feasible to keep running back to a mocap stage. We shot Thomas Hayden Church for some shots and when he wasn't available we had one of our animators, Scott Fritts, act out the scene to use as animation reference.

Hello Spencer!
First of all I'm glad to see a fresh new 'Meet the Artist'-session, and even somebody who about who's work is spoken allot these days!

I finally got to see Matrix Revolutions last week, and I must say that this film gave me the (rare!) feel of "yeah, it's really really cool that I do something in VFX in my sparetime"! I think the climatic battle for Zion is one of the best VFX sequences I've seen in a long time!

I must say I enjoyed the questions by Bonedaddy and your answers - they contained very interesting information! (as for the rest of this thread - ofcourse! )

1. Some years ago I red the VFX-article about Matrix Revolutions here on CGTalk (link), and it contained interesting information about the 'superpunch'. I was impressed by the fact it was all hand-animated and full 3D, while it seemed to be possible in a much easier way (cameramapping the head-plate onto geometry, getting the punch into that and some clever compositing, instead of a total 3D-head-recreation with cyberscanning etc). I always thought it had something to do with complex cameramoves, but when I saw the full scene I did still not understand why this all was made so complex (as the move wasn't that more complex then the other bullettime-shots). Did the choice have something to do with the watersplashes, or do I just simply get the complete method wrong?
2. On the Zion-battle in Matrix Revolutions: was you being responsible for some of the animation? (if so, first of all, congratulations ) I wonder how you worked on the 'robot-creatures' where the soldiers are hangin in. Was the scene first completely prevized by you animators, so that the motiondevices where the guys hung on, on set (in a greenscreen stage?) could be lined up? Or are most of the characters you see digital doubles? As this must have been quite a complex scene, as in some of the shots 15+ characters seem to be placed in those devices, which are all moving and fighting...! (I'm sorry if you think this is a unappropiate question to an animator, as this might be more the compositing-side of the story... )
3. You said before that you always have to do things effective, so mainly you'll especially want to be applying the directors vision the most easy. This way the disicions in the process seem to become more and more getting to be postponed to the compositing-suites and editing rooms. For instance the way relighting is these days often possible in a very far stadium after rendering, so that the director can easily change things on the 'final' frame. Do you think animation will also become more judged in the context of the final shot - so the workflow will allow very late changes on the animation? Or do you think that's just not nescessary?
4. When you look back on the old days (of stopmotion): do you like these hyper-technical than those? Do you think you got another workflow in animation because of your previous stop-motion experience?

I'll be watching SM3 very soon, and I'm sure I'm gonna enjoy your work on it! Thanks for your time, and all the best on your upcoming work!
-Gijs

Hi CB_3D,
Thanks, I'm glad you enjoy our work. Since I'm not a digital modeler I asked Koji Morihiro (rigger/animator extrordinare) to help me answer your question. Here's his reply;
"the wireframes presented in making-of's actually are the base polygonal model that the modelers sculpted.
the subdivisioned models only exist in rendering times, so no one really sees them in wireframes and they'll be too dense to tell what is going on even if you can see them."

I hope that answers your question, let me know if there's anything else you want to know.

thanks,
Spencer

Hi modelviz,
Thanks for the kind words, I'm happy to be able to chat with all of you. In answer to your question about sand simulation I must say that I agree with you, it's difficult for me to imagine how it was done too! I say that because it was the incredible talents of the sand fx team that pulled that off. I was mainly responsible for supervising the physical performance(body language) of the animated characters so for me to try and explain what the sand guys did would not do it justice. There is a great article on this site where our sand experts Ken Hahn, Doug Bloom, Jonathan Cohen and Chris Allen discuss the details of their art.
I can talk about how we (character animation) and the sand team interacted. Both of our teams worked very closely to create the Sandman in all his forms. Usually the process is much more linear, meaning that we animate the characters, then when we are done we give it to the fx team and they do their thing. This generally works fine for cloth and hair fx but for Sandman the movement of sand and it's interaction with the character required a lot of back and forth between our two departments.
Your second question, I believe, is about the "goo" that wraps around Spidey and stains his suit black. That was also a combination of keyframe and fx animation. One of your fellow posters (bonedaddy) asked a similar question earlier, here's my response that should answer your question as well;
The process was mostly hand keyframed. We tried doing it procedurally in the beginning but we found out quickly that it was going to be tough to get the organic quality that Sam was looking for solely with that technique. The basic rig was mostly spline IK for the tenticles (we called them "pods"). Koji Morihiro and Ryan Laney set up a system of maya shelf buttons that we could use to import pods and build the rig on a per shot basis. The pods included attributes for bulging and pulsing that could also be keyframed. After the keyframe performance was done the fx team added more layers of procedural goo to give it more organic detail.
The goo was a fairly specialized task within the character team. The animators that really excelled at it were the ones with strong technical skills as well as great character animation skills. Koji really helped set the standard for this character. He's a great rigger as well as a great animator.

Thanks,
Spencer

hello mr.cook

what is your wishlist for animation rigs? is there anything that you would want in a rig that is really hard to ahieve, or impossible to achieve?