Polar Express and the CGI Uncanny Valley
Dark Horizons has an interview up with Tom Hanks, who discusses his new Christmas movie, THE POLAR EXPRESS, which is coming out tihs month. Although marred by obvious typographical errors (enough to rival those you find on this site!), the interview is interesting and informative.
We learn that Hanks, who has read the book to his children for years, purchased the films rights himself and proposed the idea of making a movie to Robert Zemeckis, who wrote and directed the film. Hanks was concerned about whether the fantasy world illustrated in Chris Van Allsburg’s book could even by realized on film, but computer graphics and stop-motion came to the rescue.
This decision, inevitably, rings bells of alarm in some quarters, from people afraid that technology is taking over (and ruining) movies. However, Hanks, a fan of film technology like stop-motion (he enthusiastically presented Ray Harryhausen with a life-time achievement Academy Award a few years back) thinks the process opens up new avenues for actors to express themselves, thanks to motion-capture technology, which allows actors to give performacnes that are captured in the computer and rendered with three-dimensional-looking graphics:
“If Meryl Streep can play the greatest Ghenis Khan in history, better than anyone else can play Genhis Khan, Meryl streep can play Genghis Khan. And if James Earl Jones can play the greatest Mickey Rooney i The Mickey Rooney Story, James Earl Jones can now play Mickey rooney in the Mickey Rooney Story. It’s an extraordinary opportunity for actors to no longer be limited by size, weight, colour of hair, gender or race, which is actually really great news.”
Of course, the downside of computer-generated imagery is that, although it works fine at rendering objects and even fantasy creatures, it can yield bizarre, almost distrubing results when seeking to simulate human facial expression.
There was an interesting article in Salon.com awhile back that addressed this issue very well (I looked but couldn’t find the article in their achives). Although discussing videogames, the basic point relates to films as well: increasingly realistic computer-generated depictions of human faces are pleasing up to a point, and when they go past that point, they become unpleasant.
The problem seems to be that we accept a certain amount of lifelessness in artistic renderings like conventional animation, but when the depiction looks more realistic, that lifeless quality stands out much more clearly, and we react on some visceral level as if contronting an emotionaless zombie.
It’s for this reason that I have never been enamored of Pixar’s depiction of human characters. Their films are wonderful when portraying dolls (in TOY STORY), monsters (in MONSTER, INC), and fish (in FINDING NEMO), but everytime the focus shifts to the human characters, the limitations of CGI are sadly exposed. (The films succeed in spite of this because they are so well-written and funny that they engender a sense of good will form the audience, who then overlook the shortcomings.)
I’m looking forward to seeing how THE POLAR EXPRESS turns out. But judging from the trailer, right now I do not expect it to be an overwhelmingly heart-warming experience. As good as the backgrounds look, that robotic CGI-rendering of Tom Hanks face is not something I find the least bit endearing. My hope right now rests on the fact that movie previews often are forced to use early test footage because the final rendering is not ready in time, so maybe the actual film footage will breath some life into that lifeless zombie.
UPDATE: It appears my fears about the CGI problem may be true, even though POLAR EXPRESS is generally being favorably reviewed. The film so far has received an 80% favorable rating at Rotten Tomatoes.com, and Duane Byrge at Hollywood Reporter writes, “The film is not sheer wizardly; it also has heart.”
However, David Rooney at Variety sounds as if he must have been reading my mind when he wrote, “The Polar Express may succeed via the motion-capture process in replicating human movement by digitalizing the performances of live actors, but it fails to capture the subtlety of facial expressions or to fabricate sympathetic, evocative figures.”