So yesterday I had the afternoon free, moseyed on over to a multiplex and sat down to watch Step Up 3. This naturally means that the first question coming to mind is “Jay, you’re a sex-crazed adrenaline junkie who concentrates on action movies, what in the blue fuck were you doing seeing a cheesy dance movie sequel? Did you just need somewhere to cry over the lost potential of The Expendables?”
Valid question. The answer being that SU3D is the first movie to come out recently that was A) shot in 3D instead of converted, B) has a lot of body movement throughout by the performers, and C) isn’t a CGI-laden effects fest. That made it a workable (I refuse to use the term ideal) platform to see how modern 3D shows the way human bodies move. Kind of important for one who does a lot of work with fight scenes, no?
Not to say that I haven’t seen 3D fight scenes lately. Beowulf back in 2007, Avatar last year, and the crappy 3-D add-ons of Clash of The Titans this spring all had their share of combat. But they also were loaded with CGI effects in the process. No real way to tell what was a human in front of a 3D camera and what was a tennis ball on a stick slathered in CGI that never saw a 3D camera. Here I was reasonably sure I was watching human beings.
Early on I started noticing a particular kind of “body blur” during certain movements. At first I thought it was related to performance speed, but it would happen regardless of whether the move was done at Western or Eastern combat speed.
(Once again I start making up words and have to explain myself. Pre-The Matrix, fight scenes in western movies usually were performed about 3/4 to 4/5ths the speed of fight scenes in an Asian film. The whys and wherefores are enough for a book, let alone another article. Bottom line: it’s my term for the speed of fight scenes. Eastern is a bit faster than Western. Good to go? Roger.)
Anyways, speed wasn’t the sole culprit of the 3D body blur I was seeing in various dance sequences. But the more I watched, the more I saw it was a combination of speed and crossing planes.
And once again I have to explain what the hell I’m talking about. Fuck the ellipses, I’ll just keep going. I’ve never gotten to see any of the early 3D movies from the 50′s, and was really too young to catch much, if any of the big 3D features in the 80′s when I was a kid, so I can’t speak from experience. But in a nutshell, 3D works on showing different planes. Early or crappy 3D mostly just differentiates between foreground and background. A picture of 3 people in front of a wall would only show distinction between 2 planes: the wall (Background) and the people (foreground). As 3D gets better, it establishes more planes.
What modern 3D appears to be coming close to doing is establishing a plane for each and every object on screen, making the illusion of depth greater. (I’m far too lazy to look up tech releases, but this is what I’m seeing and guessing). Which means for the picture I mentioned earlier, each person would have their own plane. 3D body blur happens when somebody moves at speed and either crosses different planes, or occupies more than one plane at once. This is most striking with the money shots that try to make people or things jump out into the audience. A body moving perpendicular to the camera (towards the audience) can have their hands, arms, head, torso, and legs all occupying different and/or multiple planes. It’s between these planes that 3D body blur kicks in. Human eyes aren’t fooled by the multiple planes, and don’t “fill in the gaps” missing in the image.
I’m wildly guessing here, but I believe that this doesn’t affect CGI objects as much as it does real objects because modern CGI designs are rendered 3-dimensionally as a matter of course. Because it begins and ends as an image, CGI winds up being more adaptable than flesh-and-blood performers when it comes to such effects.
I don’t doubt that camera techs are already coming up with new ideas to compensate. Off the top of my head I’m theorizing some sort of dish-shaped image collector to get the POV of every point in the theater where an audience member can be. But that leaves the question of how to choreograph fights for 3D in the meantime?
Since I don’t have anything resembling the equipment for trial and error, I can only speculate, but here goes:
- Avoid Bourne Editing (excessively rapid cuts during a scene) at all costs. Fortunately this is now common knowledge among directors and editors, as it has a habit of causing nausea and headaches in viewers.
- Keep the framing of fights as lateral as possible, moving across the camera’s POV rather than toward or away from it.
- Save perpendicular action (toward or away from the camera) for “Money shots.”
That’s about all I have for now on the topic. It does mean I’ll be keeping my eye out for upcoming 3D releases. Particularly ones that put the emphasis on real-world movement than CGI.
Oh Gods, did I just come up with a good reason to see Jackass 3D? Help.
Valid question. The answer being that SU3D is the first movie to come out recently that was A) shot in 3D instead of converted, B) has a lot of body movement throughout by the performers, and C) isn’t a CGI-laden effects fest. That made it a workable (I refuse to use the term ideal) platform to see how modern 3D shows the way human bodies move. Kind of important for one who does a lot of work with fight scenes, no?
Not to say that I haven’t seen 3D fight scenes lately. Beowulf back in 2007, Avatar last year, and the crappy 3-D add-ons of Clash of The Titans this spring all had their share of combat. But they also were loaded with CGI effects in the process. No real way to tell what was a human in front of a 3D camera and what was a tennis ball on a stick slathered in CGI that never saw a 3D camera. Here I was reasonably sure I was watching human beings.
Early on I started noticing a particular kind of “body blur” during certain movements. At first I thought it was related to performance speed, but it would happen regardless of whether the move was done at Western or Eastern combat speed.
(Once again I start making up words and have to explain myself. Pre-The Matrix, fight scenes in western movies usually were performed about 3/4 to 4/5ths the speed of fight scenes in an Asian film. The whys and wherefores are enough for a book, let alone another article. Bottom line: it’s my term for the speed of fight scenes. Eastern is a bit faster than Western. Good to go? Roger.)
Anyways, speed wasn’t the sole culprit of the 3D body blur I was seeing in various dance sequences. But the more I watched, the more I saw it was a combination of speed and crossing planes.
And once again I have to explain what the hell I’m talking about. Fuck the ellipses, I’ll just keep going. I’ve never gotten to see any of the early 3D movies from the 50′s, and was really too young to catch much, if any of the big 3D features in the 80′s when I was a kid, so I can’t speak from experience. But in a nutshell, 3D works on showing different planes. Early or crappy 3D mostly just differentiates between foreground and background. A picture of 3 people in front of a wall would only show distinction between 2 planes: the wall (Background) and the people (foreground). As 3D gets better, it establishes more planes.
What modern 3D appears to be coming close to doing is establishing a plane for each and every object on screen, making the illusion of depth greater. (I’m far too lazy to look up tech releases, but this is what I’m seeing and guessing). Which means for the picture I mentioned earlier, each person would have their own plane. 3D body blur happens when somebody moves at speed and either crosses different planes, or occupies more than one plane at once. This is most striking with the money shots that try to make people or things jump out into the audience. A body moving perpendicular to the camera (towards the audience) can have their hands, arms, head, torso, and legs all occupying different and/or multiple planes. It’s between these planes that 3D body blur kicks in. Human eyes aren’t fooled by the multiple planes, and don’t “fill in the gaps” missing in the image.
I’m wildly guessing here, but I believe that this doesn’t affect CGI objects as much as it does real objects because modern CGI designs are rendered 3-dimensionally as a matter of course. Because it begins and ends as an image, CGI winds up being more adaptable than flesh-and-blood performers when it comes to such effects.
I don’t doubt that camera techs are already coming up with new ideas to compensate. Off the top of my head I’m theorizing some sort of dish-shaped image collector to get the POV of every point in the theater where an audience member can be. But that leaves the question of how to choreograph fights for 3D in the meantime?
Since I don’t have anything resembling the equipment for trial and error, I can only speculate, but here goes:
- Avoid Bourne Editing (excessively rapid cuts during a scene) at all costs. Fortunately this is now common knowledge among directors and editors, as it has a habit of causing nausea and headaches in viewers.
- Keep the framing of fights as lateral as possible, moving across the camera’s POV rather than toward or away from it.
- Save perpendicular action (toward or away from the camera) for “Money shots.”
That’s about all I have for now on the topic. It does mean I’ll be keeping my eye out for upcoming 3D releases. Particularly ones that put the emphasis on real-world movement than CGI.
Oh Gods, did I just come up with a good reason to see Jackass 3D? Help.
RSS Feed
