[Timo]

The Problem with FX

Now that the bar on special effects has been raised so high, is it impossible to clear?

July 24 � _Three years ago, for the first time in his career, Ang Lee directed a film that included the use of visual effects. It wasn�t groundbreaking stuff. �Crouching Tiger, Hidden Dragon,� as everyone knows by now, was a kung-fu fantasy in which actors sailed through the air with the aid of wires�a familiar technique in Asian cinema known as �wire fu.�
_
LEE�S EXECUTION of wire fu was solid but not remarkable. The actors lurched a bit when they took off and hung somewhat limply in the air while they flew. The presence of wires, which were deleted digitally in postproduction, was obvious. But here�s the thing: no one cared. The flying in �Crouching Tiger� may have struck American viewers as silly at first, but soon they were too engrossed in the movie to fuss.

Earlier this summer, Ang Lee released a film that included extensive use of visual effects. Unlike �Crouching Tiger,� this was groundbreaking stuff. �The Hulk� featured a muscle-bound 15-foot green monster that was entirely computer generated and had to deliver, along with the usual range of smashing theatrics, a full, emotional performance. But the creature itself wasn�t even the greatest challenge facing Industrial Light & Magic, the visual-effects house handling the task. In the movie, the Hulk had to appear in broad daylight, in our world, in the context of other humans. Nothing like it had ever been tried before. (Gollum of �Lord of the Rings,� the most successful computer-generated, or CG, character to date, didn�t have to look plausible in the middle of San Francisco.) In the visual-effects community, ILM�s Hulk was seen as a major achievement: the life in the creature�s eyes, the way light played naturally off its skin, its synthesis into its surroundings, all were deemed first-rate. Film critics, however, panned not only the movie but ILM�s work. The monster didn�t look real. Case closed. Moviegoers must�ve agreed, because after a huge opening weekend, �The Hulk� died at the box office.

What lesson, then, is Ang Lee supposed to have learned from his two experiences? He made one movie with rudimentary FX (jargon for special effects) and won four Oscars. Then he made another movie with world-class visual effects and got hammered. It�d be easy to conclude that the lesson is this: visual effects only work if the story works. But while that�s often true, it can�t always be the case. If an actor can give a great performance in an otherwise mediocre movie�as Johnny Depp just did in �Pirates of the Caribbean��why can�t a lousy movie have great visual effects? And no reasonable person would argue that the effects in �Crouching Tiger� were better than those in �The Hulk.� They weren�t. There�s not even a comparison. Here�s the real lesson for Lee, for ILM, and for any studio planning its next big action spectacular: the FX bar has been raised so high�call it an FX arms race�that it�s almost impossible to clear it.

One of the great myths going around Hollywood now is that, with CGI, anything is possible. Photorealistic humans! Emotional performances! Virtual reality! The truth is, all of us still know CGI when we see it. Even the best work has an indefinable �digitalness� to it that FX artists haven�t yet, and may never, overcome. Large objects don�t seem �heavy.� Textures, such as clothing or skin, seem too smooth. Movement is either herky-jerky or implausibly fluid. And yet, at the same time, CGI has come close enough to complete verisimilitude that audiences now expect it to go all the way. So when we see it onscreen, we have a strict pass/fail barometer. We either think, �Wow, that looks pretty real��which is rare�or we think, �Forget it, that looks fake.� Curiously, we are evaluating something we know to be synthetic on the basis of how near it comes to a goal it can�t reach. Is it any wonder then that, more often than not, CGI fails? The game is rigged against it.

It wasn�t always like this. Before the visual-effects industry was transformed by CGI in the early 1990s�thanks to films like James Cameron�s �Terminator 2� and Steven Spielberg�s �Jurassic Park��the bar was much lower. The original �Star Wars� trilogy featured revolutionary special effects�but no one mistook what they were seeing for reality. Chewbacca looked like a tall man in a furry suit. The spacecraft looked like plastic models floating in a dark room. The story was marvelous, of course, but the same story today with similar production values would get laughed out of the theater. We weren�t dumber back then, or less observant. We just had lower expectations. The original Christopher Reeve �Superman� movie is another example: the flying in that movie wouldn�t fly today. But it was good enough for 1978.

These days, there�s a new �Superman� or �Star Wars� every weekend. With each blockbuster movie, studios are trying desperately to show audiences something they�ve never seen before. But in most cases, they shouldn�t bother. For one thing, a great story is cheaper than great visual effects. And if you do spend the dollars for great CGI, there�s no guarantee you�ll even get credit for it. When it comes to visual effects, there�s no such thing as an A for effort. Just ask the folks at ILM.
_ _ _ _
Devin Gordon is a general editor for Newsweek covering pop culture.

http://www.msnbc.com/news/943726.asp

Comments?

[daimoni]

.

[Logic]

To me it seems lately like filmproducers are adding in scenes into movies just for the sake of showing off their skills in creating visual effects. One example that was terrible was the "parachute-surfing" scene in the latest Bond movie. It was a completly useless scene, it is still today impossible to actually do such a scene without everyone noticing it and it just looked terrible.

I've been asking myself this question for a long time. Why do producers add in scenes just for the sake of it, and why does it look like they actually write scenes around something they want to do in respect to visual effects instead of writing a good story and using visual effects where needed?

I hope you get the point I'm trying to get out there.

[AB^2=BCxAC]

Except for the plastic models in a dark room (that was the Black Hole and Battlestar Galactica-- Star wars was a bit better IMHO), the guy has the argument made solid. But it's like saying fast food is fatty, it's so obvious Hollywood doesn't care about story anymore for 95% of its efforts.

[Nonsuch]

Originally posted by daimoni:
The author totally misses the point of what Ang Lee was trying to do with Crouching Tiger. If he expected nitpicking 'realism'... then he was looking in the wrong place (a stylized fairy tale set in a culture different than his own).

I didn't get that at all. He's simply pointing out that Hong Kong-style wire fu is primitive compared the special effects American viewers are used to. And he's right. People complained that Spider-Man looked to light and wobbly -- too digital -- as he swung around Manhattan. It stands to reason they might complain about the lurching weightlessness of the characters in CT, HD--but they didn't, mostly because the spectacle is so beautiful in and of itself that the lack or realism doesn't hurt it; in fact, it helps. We are learning -- slowly and reluctantly -- that there is no substitute for seeing real human beings in motion.

[Timo]

I thought this part was particularly of interest:

The truth is, all of us still know CGI when we see it. Even the best work has an indefinable �digitalness� to it that FX artists haven�t yet, and may never, overcome. Large objects don�t seem �heavy.� Textures, such as clothing or skin, seem too smooth. Movement is either herky-jerky or implausibly fluid. And yet, at the same time, CGI has come close enough to complete verisimilitude that audiences now expect it to go all the way. So when we see it onscreen, we have a strict pass/fail barometer.

Why don't CGI objects seem "heavy"?

[Nonsuch]

Originally posted by Timo:
I thought this part was particularly of interest:

Why don't CGI objects seem "heavy"?

It is an interesting question. I would guess that either the animators simply don't have the training to animate heavy objects/people realistically, or the software they use doesn't afford them the capability. Regardless, I think it comes down to the animation itself, the way the object is moved on the screen, rather than anything technical like the lighting or the texture rendering.

[Adam Betts]

ILM = Crap

End of discussion.

[daimoni]

.

[scottiB]

The author's original comparison between Ang Lee films is weak: there is nothing "CGI" about CT, HD except the removal of the wires. Current "CGI" techniques wouldn't have made the wire-fu execution any better.

The topic the author seems to be meandering about in fulfilling his column is "suspension of disbelief." While CT, HD does have a compelling story, that it's set "a long time ago in a culture far, far away" is what allows us to ignore the somewhat imprecise wire work (when the characters walk along and among the tree boughs). The same with Ray Harryhausen's monsters in the Sinbad movies: the films are set in storybook legend, and we can ignore the stop-motion lack of fluidity.

Regarding the Marvel Comic movies, there's has been criticism (and I agree) that some of the CG characters aren't "life like." It doesn't matter to me (or hinder my enjoyment of the movie) because I consider them to be modern Sinbad flicks, but it seems it does to bother others (I don't think the reason The Hulk failed at the BO was that the Hulk wasn't life like as much as the weak script. Also there are too many Marvel motion pictures being released in a short time span).

A difficulty, I would think, in animating "weight" is to have a real world example as a reference. With a Spider-man or a Hulk, that's tough to do. That's why they're comic books.

Really good CGI is the stuff that you're (or at least I) am unaware: the atmospheric or background stuff (like the propeller of the first downed helicopter in Black Hawk Down as it strikes the ground and breaks apart).