Interview With Michael Arias

Michael Arias works for the Softimage special projects team and talks about creating The Animatrix movies, the industry and the fusion of 2D & 3D techniques.

September, 26th, 2003, by Raffael Dickreuter, Bernard Lebel, Will Mendez

Tell us a bit about your background and how you got started in the 3d industry
It always seems to me that my career’s taken many bizarre turns along the way. I started out in the film industry – in 1987 or ’88 – at a company called Dream Quest Images – later reincarnated as Disney’s Secret Lab, and now defunct, as far as I know. I was just looking for something that would make use of my electronics skills and keep me busy while my band sorted out its various personnel problems. After a little while there, I ended up helping debug Dream Quest’s new motion control system, and then being chosen to work on the stages as an assistant, since I was one of the few people who knew their way around the system. Those were great days for the effects industry – it was before the advent of much CG and people were doing tremendous things with motion control, optical effects, miniatures, pyro, whatever. And the studio was a great place for me – still 19 or 20 – a big tinkertoy factory run by car nuts and mad bikers. At DQ I got to work on some great films too – THE ABYSS, TOTAL RECALL, and some others.

working on James Cameron’s The Abyss.

After a couple years there, I moved back to the East Coast and promptly got recommended to Doug Trumbull for his BACK TO THE FUTURE, THE RIDE attraction for Universal. Another great experience – Doug was – IS – such an inspiring figure. For me and the other younger crew, including John Gaeta, now VFX Supervisor on the MATRIX films, Doug was so generous with his knowledge; such a very warm and receptive and articulate and creative guy.

It happened that our optical composites were being done by a Japanese company called Imagica. BTTF was all Omnimax so it required large format opticals, and Imagica was one of the few places in the world where one could do high-quality 15-perf opticals. And because I spoke some Japanese (from having studied Japanese in university), I ended up spending a great deal of time with the Imagica folks when they were in town. That, combined with the fact that our miniature crew was also largely Japanese, left me with a standing invitation to come visit Japan.

I ended up finally going with Doug to the Osaka expo (in ’90 I guess) and, though I’d seen the CG in THE ABYSS and a couple other shows by that time, the stuff I saw in Japan just blew me away. That, more than anything else, was what convinced me that the future of filmmaking was CG. I ended up working at Imagica for a year after that, still doing motion control camera work. And then I got the chance to direct a short “ridefilm” at Sega, for an 8-seater hydraulic motion base they had. Of course I had no real experience with 3d except for what I’d managed to learn from books and a borrowed copy of 3DStudio (rev. 1!). But the Sega folks saw this as an interesting opportunity to build up their CG team, still using Iris4D workstations at the time. And by the time I was onto the project they had chosen Softimage|3D, then called Softimage Creative Environment, version 2.4 I think. Perfect for a newbie like me. The film I did at Sega, MEGALOPOLICE TOKYO CITY BATTLE was shown in SIGGRAPH’s ’93 Electronic Theater. By today’s standards it’s pretty goofy but for the time it was quite ambitious. Insane really, considering that none of the team had any real CG experience.

After that I moved back to New York to team up with some friends, Randy Balsmeyer and Mimi Everett (their company is now called Big Film Design – BFD), to start a little CG production house, called “Syzygy Digital Cinema”. Through their work in film titles and design we ended up with some great clients: David Cronenberg, The Coens Brothers, Spike Lee, Jim Jarmusch. But New York wasn’t quite ready for the CG business yet and because of our feature-film focus we couldn’t cash in on New York’s commercial business (which has since all but dried up). But it was a fun two years.

And by that time, I was ready for something new. I had really been thinking for a while that to go any further in CG I’d need to program more. I’d written a package called M/CAD for Sega and Doug’s motion base programming, and that had really whet my appetite for coding. And as my time at Syzygy started getting short, Softimage seemed like a natural fit. I’d made some close friends at SI; it was (and still is) a fairly tight-knit group. And I loved Montreal (having only been there in the Summer, thus far). David Morin, then director of Special Projects, made me a very nice offer that included lots of travel to Japan, and so I joined up. This was still the “good old days” when workstations and software were costing 40, 50, 60 thousand dollars, and up, so it really was quite different than it is now that everyone’s tightening their belts.

I think I’ve been with Softimage for eight or nine years now. I’ve now actually outlived all of the folks who hired me. And I spend so much time away from Montreal that most of my co-workers only know me from email.

Can you tell us what is the role of the special projects team 
We used to joke that Special Projects was where Softimage put people capable of doing everything and those capable of doing nothing. Not sure about that, but SP does seem to have always attracted people with a jack-of-all-trades sensibility. At the time I joined I think the entire team was composed of people with heavy production experience and even now I think we tend to gravitate towards hands-on work. Some of us, like myself, do a bit of programming. Others are experts in technical animation, scripting, or games development, for example.

Our focus is fairly short term, by necessity. Even our software projects are limited in scope by our commitment to a particular client, or by how much we have to work on many things at once. Not at all like how the folks in R&D; spend their time. That said, though the R&D; folks may not have the practical knowledge or the user’s outlook, they definitely have to deal with the software at a much lower level, and their knowledge of its workings, and of the mathematics and methods of 3D, are way beyond that of any of us in Special Projects (or at least me). The smartest people I know are all in R&D; at Softimage.

Did you always work for the Special Projects team at Softimage, or what were your previous positions?
I’ve been with SP since I joined Softimage. Actually I think I’m the only original member left at this point. Kind of a dinosaur, really. Someone has to come from Montreal periodically to clean off the cobwebs. Change my fluids, filters, and hoses, that kind of thing.

What do you do in your spare time?
Play with little kids. My kids, that is. Also, I’ve been biking a lot lately. Biking seems to be the latest fad to hit the traditional animation industry in Japan these days. I just did a 120km ride with Katsuhiro Otomo (director of AKIRA and the upcoming STEAMBOY). That almost killed me. Everyone else’s bicycle cost 5 times what mine did and weighed half as much. When I showed up at the rendezvous, the first thing the others did was pump up my tires and offer to strip my bike of all the extraneous bits (kickstand, baby-seat rack, etc.).

Working in the industry do you find that projects are becoming more technical than artistic?
I’m not sure if you mean the CG industry or the film industry in general.
If you mean the CG industry, I’d have to say “not really”. Software is (slowly) getting easier to use, and this means that CG artists, in some positions anyway, require less technical knowledge. That’s opening the field up to more talented artists with a broader range of talents. That’s good.

Filmmaking has always been a mixture of the artistic and technical, but I think slowly filmmaking too, particularly with the advent of non-linear editing and digital cinema, is becoming more open to artists who might have otherwise been inclined to pursue more direct means of expression. It’s easier and cheaper to make movies these days. At least, I think one could say, the entry cost is much lower. You can shoot a movie digitally, and mix and edit it all on a home computer, achieving respectable quality at the same time. That kind of thing was unthinkable just a few years ago.

What is your view about the current situation in the industry?
I think, on the one hand, the CG software industry is in a very tough spot, even while these are good days to be a CG animator. The film industry, like always, suffers from a lack of good ideas. There’s just so few good movies made.

What are the biggest differences between the Asian market and the Western 3D market?
I think the biggest difference is in the size of the market, in general, at least for artists in the film industry. Though it wasn’t always the case, the Japanese film industry is miniscule compared to that of the US. And movie budgets here reflect that. And lower budgets mean typically mean fewer effects, hence less CG.

On the other hand I think the game industry, though it’s definitely seen better days, still offers some interesting work, both for artists and software developers.

How has the localization of XSI help improved the Japanese market? 
Well, the people here at Studio4ºC (Studio Four Degrees) started using the Japanese version then day they got it and haven’t switched back to English. I have no idea if the introduction of a region-specific interface has helped sales here, but I can’t imagine otherwise.

You’ve worked closely in the Animatrix project. What can you tell us about that?
Andy and Larry Wachowski, directors of the MATRIX films, and their producer, Joel Silver, contacted me through their VFX supervisor, John Gaeta, an old friend who I’d kept in touch with over the years, and who knew I was working in Japan. We all met in Tokyo, and then in LA a couple of times, and after we’d talked about their idea for an anime “dream-team” project a couple of times, they asked me to produce the project for them.

I had very little “production” experience. Nothing really, except for once having acted simultaneously as CG director and co-producer of a feature-film pilot (TEKKON KINKREET, seen at SIGGRAPH 2000). But I felt good about the Wachowskis and the folks at Silver Pictures, and, more importantly, I had great partners in Japan: Eiko Tanaka, president and producer at Studio4ºC (where we ended up making much of ANIMATRIX), and Hiroaki Takeuchi. Tanaka stayed pretty close to her studio, while Takeuchi dealt with a couple of our other studios, and oversaw legal and contractual issues. Regardless, it was an enormous responsibility, and it totally dominated my life for the three years I was on it. I think in many ways it was more complex than producing a feature might have been, simply because we had so many teams running in parallel, and each director aspired to make their own “mini”-feature. Traditional animation in Japan has so much to do with personalities: some directors and animators require a great deal of hand-holding, while others are very independent. Fortunately, because I had worked a great deal in the animation business here already, I was on a “first-name” basis with many of the staff of our various episodes.

I really had to draw on a great deal of experience that had sat unused in the background while I’d been pursuing software development. Everything I’d learned until this point: a brief career in recording studios, composing music and doing sound effects for short films in college, having my own company, working in special effects. It was a great chance to exercise some dormant (or damaged) brain cells.

Honestly, though I didn’t get involved in the computer graphics aspect of the films as much as I’d have liked to, the most enjoyable part of the process for me was post-production. None of our directors came over for the dubbing sessions and, even for the fellows who made it to their final mixes and met with the composers, I was able to act much more as a collaborator than most producers are. This was not only because of the language barriers, but also because the post-production of animated films in the US is so different from the Japanese way.

Without question, my favorite single day on the show was towards the very end, when I got to record “walla” (crowd noise and background voices) for the battle scenes in SECOND RENAISSANCE. They had me in a booth screaming at the top of my lungs for hours – enormous stress relief. I’m especially proud of my soldier, begging for mercy while having his arms ripped off by a Sentinel (“oh God, no, please, God, no, AAAARRGGGHHHH!!!!”).

To what extent XSI was used in the overall project?
I think all of the episodes done at Studio4ºC were done using XSI, though there were some models built with Softimage|3D. You have to remember that we started ANIMATRIX three years ago, and there were some major gaps in the modeling functionality then. By the end, when we were working on BEYOND, we were doing everything with XSI version 2.0.

Kawajiri’s PROGRAM had a couple of CG shots done using Softimage|3D.
CG elements for Peter Chung’s episode MATRICULATED were done by a couple shops in Seoul using mainly Maya and Max. Of course, the only full CG episode, Square’s THE FINAL FLIGHT OF THE OSIRIS, was done using Maya and Square’s in-house renderer.

As a producer, I wasn’t necessarily in the best position to deal with software choices, but I think all of the Japanese animation houses saw a clear advantage in using Softimage software. Being on close terms with a software provider definitely gives a production an edge, particularly when technical challenges arise. Just the fact that I was writing the XSI Toon Shaders and dropping off new versions at the studio almost daily was seen as reason enough to use XSI, particularly on SECOND RENAISSANCE and BEYOND, which contain so many hybrid 2D-3D shots.

The Second Renaissance

Why do you think was XSI the perfect choice for a project like the Animatrix?
Great set of tools useful for 2D/3D integration: the Toon Shaders, of course. But also the camera projection mapping features, lens center offsets, compositor, audio tracks, animation mixer. The Render Tree and interactive rendering with mental ray. Good Japanese documentation and great local support.

Were there any special techniques used in combining 2D and 3D artwork
No rocket science really, but I think we did end up doing some wonderful shots with 3d characters (Toon rendered) and hybrid 2D-3D backgrounds. Studio4ºC has really refined the techniques involved in “perspective mapping” – projecting hand-painted artwork onto 3D geometry to camera movement other than just 2D panning and zooming.

You’ve worked on the development of the XSI toon shader, right? What can you tell us about that?
One of the first things I wrote when I started at Softimage was a very simple shader to do two-tone rendering – a poor-man’s cel shader if you like. It was quite primitive. I think most shader-writers start with similar projects. But my boss at the time, David Morin, thought that it might be useful for a project that was then being worked on by MTV’s digital team, DTV, and he put me in touch with their director, Myles Tanaka. Myles asked me to check out the work they were doing on a television pilot called “The Cathy Sorbo Show” (or something similar). The show involved cartoon-rendering a motion-captured performance of a talk-show host. The original idea was to do everything in real-time, but my shader was eventually used to offline render all the characters.

I kept working on the shader after that. Warner Bros. had me work with their rendering guy for the Duck Dodgers thing they did with Michael Jordan, showing him the techniques I was using to get my ink lines. They ended up incorporating the same techniques into a Renderman-based pipeline.

The big advances came when Softimage put me in touch, more or less simultaneously, with Dreamworks and Hayao Miyazaki’s studio, Studio Ghibli. Dreamworks was ramping up for PRINCE OF EGYPT at the time, and Ghibli, for PRINCESS MONONOKE. I spent the better part of the next couple years writing a library of ink-and-paint shaders to their specifications, and managed to get a patent on some of the techniques I was developing at the same time. Dreamworks went on to use the shaders on some amazing shots in THE ROAD TO EL DORADO. Full 3D characters, rendered to match the traditionally animated elements – really convincing stuff. I’m quite proud of their work. I’m no longer in touch with Dreamworks, so I don’t know how theye are going about these kind of shots now, but Ghibli continues to use the latest version of the Softimage “Toon Shaders”. While producing ANIMATRIX, I rewrote them from the ground up to take advantage of XSI’s Render Tree and interactive rendering as well as some new mental ray features. We used them quite a lot on THE SECOND RENAISSANCE 1 and 2 and BEYOND.

All in all, I think I’ve been playing around with these shaders for six or seven years now. As software, they’re really not so complex. The real key to their success in the field is the feedback that I was receiving from various animation studios testing the shaders for me. Because I had regular contact with all of the key Toon Shader users while working on each new version, I was able to incorporate a ton of suggestions as well as analyze users’ reports to pinpoint bottlenecks and deficiencies in the software. Along the way I think I managed to stumble onto a couple of clever tricks as well. I taught myself programming and math and computer graphics for the most part, so anytime I write software I end up employing a “hunt-and-peck” methodology. A more experienced programmer would no doubt arrive at a better solution faster. But I learned a great deal from the clients and Softimage people I was working with, and consequently really enjoyed working on all of this.

Is there a difference between anime animators and 3D Animators, if so what is it?
I haven’t seen any 3D animators in Japan with the chops of many of the traditional animators here. 3D is a somewhat deceptive tool: you’re able to create fantastically real (or surreal) images with comparatively very little effort. Things like motion blur, depth of field, sophisticated camera movement, rigid body dynamics, particles, hair, etc. are all included right out of the box. But if you give most 3D animators a simple skeleton with minimal rigging and automation and ask for a good “sad” pose, for example, or some athletic running and jumping, without relying on the technical animator’s bag of tricks, I think you’d be quite disappointed with the results. So when looking at various 3D work, you might see very evocative or convincing still imagery, but when you evaluate it for animation, you’ll often be disappointed by performances that are below the level of even some very rudimentary traditional work. This isn’t to say that there’s no bad trad. animation here – there’s tons of it – just, generally speaking, the training and experience and talent of traditional animators is of a higher level. That said, there’s an extraordinary shortage of good traditional animators, while there’s a phenomenal glut of 3D artists. As far as I can tell, the situation is similar, in both regards, in other parts of the world as well.

What Do you think about the fusion of 2d and 3d techniques? Will 2d disappear?
I don’t think it will ever completely disappear, not as long as there’s a few people crazy enough to want to actually hand-draw their films, frame by frame. At the same time though, everyone here complains about the lack of good animators. There seems to have been an entire generation of artists that chose other fields – CG for instance – instead of becoming traditional animators. The youngest animators on our staff were already pros when AKIRA was being made, and that’s an old film now. And these guys have never had the chance to pass on their knowledge to the next generation of animators. The budgets stay the same or shrink, while people’s expectations get higher and deadlines get tighter. It’s tough work; can you imagine actually hand-crafting a film? There’s nothing on the same scale except for animating with clay or models.
Even in the US, traditional animation has lost enormous ground to CG, though I think a great deal of this has to do with the relative brilliance of the stuff being put out by Pixar and Blue Sky. I really do feel like ANIMATRIX proves that people will watch good animation with compelling stories.

But yes, I suppose I do feel like we’re seeing the slow decline of traditional animation. Sad, really.

What challenges would you like to take on? 
I’m currently storyboarding a feature-length animated film that I’ll be directing at Studio4ºC.