The Mark Harris fallacies

Today David dropped off an article from Entertainment Weekly by Mark Harris called "Micro-Mangling." It's about the film 300 and its visually perfectionistic, blue-screened inhuman sterility. Harris goes on to talk about directors like David Fincher, Darren Aronofsky and George Lucas and the similar problems with their films.

Reading the article was helpful because it helped me realize that I need to differentiate between perfection and performance. Let's take the example from the film Stealth that I always use—a plane whooshes by as a virtual hand held camera shakily pans from left to right in an attempt to keep the plane in frame. The problem isn't that this shot is 100% virtual or that the cloudy sunset is too perfect. The problem is that the shot is poorly performed. The virtual camerawork, however it was done, looks like someone saying "hey, look at me, I'm handheld!" This is where the idea of witholding comes in. Authentic camera work looks like someone trying to keep the camera steady. A good camera performance would have looked shaky in spite of itself. It's the same thing with the "drips on the lens" all-CGI shot from later in the film. The shot doesn't ring true because it seems to be a celebration of splatter—not inadvertant splatter that occurs in spite of a camera operator's best efforts.

That's the thing about CGI—it foregrounds performances in areas not typically thought of as performance-oriented. One thing Harris points out is that in 300, the skies look too perfect causing them to appear inconsequential. Again, this is a problem of performance. Snyder has claimed that the 300 skies are like mood rings indicating the temperament of the scene (from WIRED)—

WN: Did you go through the film scene by scene and manipulate the backdrops to match the emotional impact you were after?

Snyder: Absolutely, I would adjust the sky in every scene so it reflects the mood of my actors. It was like a big mood ring.

One of the characteristics of poor performance is "indication," what I've called "face acting" in the past. Here, a performer is trying to overtly express a certain emotion resulting in a too-broad performance. The problem with the 300 skies isn't that they are perfect, but that they indicate; they are poorly performed. A well performed sky, again withheld, might look like a beckoning storm. Or it might play ironically as a stark, cloudless gray mass. Like background music, the skies ought to serve as counterpoint, not bombast. The problem with the 300 skies isn't perfection, it's vulgarity.

The incorrect conclusion that Harris reaches is that these films are overcontrolled or over-art-directed. By implication, the solution is to add grit or grime or imperfection or randomness. This is one of the most common fallacies of computer simulation: realism = randomness, noise. The visual effects people in the CINEFEX article I quoted in an earlier blog entry also fall into this trap. But as an artistic endeavor the goal of art is not to replicate the apparent "noisiness" of the world, but to behold an emergent moment coming into being.

The common fallacy, also one that I've fallen into from time to time, is that performance value derives from specific techniques or methodologies. That speaks to the fallacious origin of the Noise film. As you recall, the film was going to be a real-time performance captured on tape until I realized the problem: this would not make the film any more dynamic or "live." A performance can soar whether it is experienced live in a theater or captured on film. Further, capturing a performance in real time does not necessarily increase its performance value. A good animator, for instance, can draw a good performance from the outside in, frame by frame, over a period of time that is anything BUT realtime. So performance value is not a result of approach, but of understanding the way performance is expressed within a particular medium. And what's making the all-CGI films look so problematic is that we have not yet come into a full understanding of how to perform skies and blue screens and virtual camera work. What I was trying to do in the "Performance Conjectures" post from earlier this month is describe some qualities of a good performance whether it involves acting, or the creation of props or using CGI.

Here's another example. When making the first Caspar movie, Dennis Muren and team tried mo cap but went back to traditional animation when they found they couldn't get the expressive results they wanted. The problem wasn't with mo cap. The problem was that people at that time didn't understand how to render an expressive performance using the technology. Since then, Andy Serkis (Lord of the Rings) and others have begun to refine the craft of mo cap performance. In the same way that film acting is different from stage acting, mo cap acting requires a certain approach to performing.

As I've said many times before, "the shot is the performance." What CGI does is give us greater control over the shot and in that way, expands the ways in which performance value is expressed.