How 3D Changes the Landscape at Sony Pictures Imageworks
ROB ENGLE: The word has been around for a long time. It means “people who take photos in 3D.” Stereography is the part of my job that involves the creative aspect: how deep do things go? Where do we emphasize depth? Where do we emphasize where people are watching? Hence, stereographer.
F&V: Were there advantages in terms of 3D in working with an all-CG film?
BUZZ HAYES: Essentially, in a CG world we deal with perfect optics and cameras. There’s no distortion. There’s no problem based on how it’s photographed. So we can do things relatively perfectly, if you will. Whereas in the physical world, the cameras themselves have limitations.
F&V: At what point did you start planning the stereo 3D elements?
BH: This is essentially [Zemeckis’s] third 3D movie, so he has concrete ideas about 3D. We sat down with him when we had the first few finals from 2D and talked with him about how he wanted to use 3D in this movie early on.
RE: Bob knew he was making a 3D movie from the outset and he planned his shots thinking about the 3D version. There are definite moments where you’ll go, “Oh yeah, Bob knew he was making a 3D movie,” because there’s interesting composition, whether it’s a spear pointing at the screen or it’s a monster right in the center of the frame, right in your face. These are moments he’s thought about.
BH: And, when you talk about stereo, it’s not just the overt stuff. There are certain compositional elements you can use in 3D to enhance the effect.
F&V: Can you give an example?
BH: Just in terms of how he’s framing the shots. If you have something in the foreground, you have a much greater sense of depth, because you have a point of reference. So in a lot of cases, there are subtle 3D depth cues, like furniture in the foreground, that give you a point of reference for the rest of the shot.
F&V: Did you do a 3D script?
RE: We think about the movie as a whole and this is where the idea of a “stereoscript” often comes in. We think about which scenes we want deep and in which scenes we want to give the audience a break and let them just take in the 3D-ness of it. For the most part, my philosophy is that an action scene or something that really screams to have depth are the moments we’ll give depth. And we’ll dial it back for the talking scenes, the expositions. The paradox is that the scenes when people are just talking are the moments when people look at the rest of the frame, when they really understand they’re in a 3D world. During the action scenes, they’re always looking at whatever is moving the farthest.
F&V: The obvious use of 3D is for things coming at you from the screen, but did you also use 3D for emotional arcs?
RE: One of the tricks we’ve played with on this film was using depth to portray the power of the character. We tune the internal depth of characters based on where they are in their power arc. For example, there are many times when Beowulf is confused or unsure about what is going on because even though his character has a lot of bravado, his character is not always in a position of power. So we play with depth.
F&V: What do you mean by "internal depth"?
RE: Internal depth is how the audience perceives the character. When they have internal depth, they look and feel round on the 3D screen; the nose appears to stick out and the eye sockets appear to have depth.
F&V: How do you tune the internal depth?
RE: We adjust the cameras. One of the primary controls we have when doing a stereoscopic film is how deep everything appears – and deep doesn’t mean only how close something is to you, it’s also how round it feels. We can control that independently. We can control how close it is to you, how close it is relative to something behind it, and its internal depth – how round it feels. As we move the cameras apart, we introduce depth; the more we pull them apart, the deeper everything feels. So because we’re working within a virtual world, we can tune the depth for an individual character.
F&V: So you aren’t working with the scene as a whole?
RE: Well, initially we do. The first thing we do is go in with a single stereo camera. Then we decide what we need to adjust. I might take a picture of the scene and then say, “You know what? This character feels a little flat.” So I might use an extra camera to give him more dimension. I might take his picture with a wider lens and with wider separation and effectively make him deeper internally.
The problem is that if you go too far with that, you start to play around people’s perception of reality. And if I did that with Buzz [sitting across from Engle in the room] – made him super-dimensional, really deep – then he would feel giant. If I photographed him with the cameras far apart, and everything else in the scene with cameras closer together, then all of sudden I’m starting to play with perception because people don’t see in that way.
F&V: How many cameras can you have in a scene with different 3D depths?
RE: You need to make sure they fit within a consistent world. For example, if I photograph this scene – Buzz sitting in the chair – with little camera separation and then go in and use a much wider separation for him, he’s going to start to look like he doesn’t fit in the chair properly. So the challenge with using multiple cameras is ensuring that you don’t mess it up. You have to be very careful. In Beowulf, in some cases, I used a hammer, and I did it because I wanted to make it really clear: hey, this is deep. This is interesting. In other cases, we went very subtle.
F&V: What would be a "hammer" case?
RE: When we wanted to emphasize the power of a character. Maybe the character is standing next to someone else, and even though it will make this character look unusual, I’m going to emphasize this character. A perfect example of something that was intentionally done is the dream sequence when Beowulf is being seduced by Grendel’s mother and he awakes inside a giant gallery hall. I explicitly chose to use wider cameras for that entire scene to emphasis the weirdness of it, the dreamlike quality. So when you watch that sequence you’ll see that he’ll feel extra deep and you’re not sure why. You don’t know that it’s a dream at first.
F&V: Now give me an example of subtle.
BH: One of the things you have to keep in mind is that our brains know when something doesn’t look right, and there are all sorts of things that can happen in a stereoscopic world to make our brains confused. It could be as subtle as a foreground object being cut off by the edge of a frame. The relative positioning of things can mess with your mind in subtle ways.
F&V: Did you have to change anything for the 3D version?
RE: In some cases, we had to redo the shot. One example is a shot, a 3D moment, when King Hrothgar throws his bundle board onto the ground, and the bundle board comes right at camera. But, the way it was composed broke the frame. So we moved the camera.
Also, Bob likes to put the camera right on the ground. In a shot with a dragon flying right into camera, all the stuff on the ground was right in your face, which took away from the dragon in the foreground. So, we scooted the camera up a little.
BH: We also had situations, for example, where, in the traditional compositing stage, to make a character feel smaller, they would essentially push it back in three-dimensional space, and in 2D it would look like the character got smaller. In 3D the eye lines of those characters, which used to be looking at each other, would be looking past each other because one of them is behind the other and not in the same plane.
It’s subtle, but as soon as you see it, you know there was a cheat. So there is a certain thought process that has to go along with these things, especially in compositing, where you use atmospheric effects such as smoke, fire, sparks whatever. It’s always a good idea to have volumetric stuff, because it’s going to look funny in 2D if it’s a one-dimensional plane of rain and everything else is dimensional.
F&V: Are there any major differences between the 2D and 3D versions?
BH: One of the biggest differences you’ll notice between Beowulf 2D and Beowulf 3D is the use of depth of field. In 2D, we add blur so you focus on the foreground character, because in 2D everything is in the same plane. But in 3D, our brains work differently. If you add blur to 3D, your brain senses that something is wrong. In the real world, I can look at you and then very quickly shift my gaze to the back wall and that’s also in focus as least as far as my brain’s concerned. And that goes all the way back to thinking about designing the backgrounds. We had to make sure there was detail in them. We couldn’t count on having them out of focus.
Bob [Zemeckis] thinks about framing, staging, all this stuff. He has this little 3D thing in his head reminding him of this.
F&V: So if you have a scene in which you know you’ll want blur in the 2D world, what do you do?
RE: Basically, the studio first produces the 2D version – our 3D has always been a complement to the 2D version of the movie and the 2D movie has always been slightly ahead in physical production time. When that’s done, we take the virtual-world assets – the textures, lighting, character animation and so on – and create a second camera that’s sometimes to the left, sometimes to the right. Sometimes we change both the left and right eye.
When we create the second camera, we’re conceptually taking another photograph; we’re rendering the other eye. Then we go in and re-composite the entire shot for both eyes, taking out things like blur and depth of field that have been composited in.
F&V: With everything in focus, do you see a lot of detail?
BH: You’re much more aware of the detail in 3D. Especially when you see Hrothgar in tight closeups. You also see tiny bits of peach fuzz across the women’s cheeks as the light hits it.
F&V: Do you see that more in the 3D version than the 2D version?
BH: Much more in stereo 3D. I think it has to do with the way our brains work. Two eyes’ worth of information lets you rectify whether that’s hair or light or what. If you have only one eye, you might pass it off as a little bit of light on the side of her face. But in 3D, we have just enough perception here and there to tell the difference between a glint of light and hair catching the light.
F&V: Why create Beowulf as a CG film in the first place rather than live-action?
RE: I’ve heard people say that Bob is interested in being able to control every nuance of his film and by using this technique – not just 3D but CG in general – he is able to exact a level of control over the final product that he can’t with traditional filmmaking and real cameras. And CG gives us a level of control for producing a quality 3D version that would be a lot harder with live action.
F&V: How much was Zemeckis involved in the process of translating the film into 3D?
BH: It’s not like we have a blank canvas. We know what he wants to see. He gives us only a certain number of choices based on where he puts the camera, for example. There’s only so much you can do with it. And, he’ll do crazy things, like you’ll see a close-up of someone and the camera pulls back through the handle of a flagon or something, these crazy impossible camera moves that you could never do in the real world and yet he does them all the time.
Rob pulled up a series of shots that represented the movie in terms of how various characters would interact in a tight environment and a wide environment, that sort of thing, so that Bob could get a sense of overall depth and what it could look like. He gave us notes on that, and that gave us our road map.
F&V: Did you run into any particular problems?
RE: What I can do creatively is determined a lot by what the technology for exhibition is capable of doing right now. For example, the geometry of the theaters and the phenomenon of ghosting. When things have a lot of separation, especially in scenes that are very contrasty, you tend to get ghosting where you see a little of the left eye image in your right eye and vice versa, because right now the exhibition technology isn’t perfect. So all these little things, all these little pieces of information about how the technology works, you need to be aware of as you compose stereo.
F&V: Did you have to do a separate DI?
RE: We are doing a separate DI, and the reason is that, if you think about the glasses, there’s a significant amount of light loss. The left eye doesn’t get 100 percent of the light. Same with the right. In fact, it gets a small fraction. Less than 50 percent of the light gets to each eye. So, as a result, the filmmakers would like to bump up their movie, make it a little brighter to compensate for the fact that these 3D filters add an extra bit of darkening.
F&V: If Beowulf had been a live-action film, would it have been possible to make it as three-dimensional?
RE: Yes, but it would have been a lot of work.
BH: One of the problems with live action for 3D now is that there aren’t a lot of people seasoned to do 3D. If you’re dealing with a crew that’s not seasoned on it, then the chances for getting material that may not work later is fairly high.
F&V: What kinds of problems would people run into?
BH: It’s very difficult to do close-ups in live-action 3D because you can’t get the lenses close enough to each other and to the subject to read as a close-up, so there are certain tricks you have to take into account.
Alfred Hitchcock’s Dial M for Murder was originally shot in 3D and he ran into that problem because the crux of the movie is dialing M and you couldn’t do a close up with the old cameras. So they built a giant telephone and gigantic finger. If you look closely at the shot, you’ll see it’s a fake finger.
And you can set up some beautiful moments, but once you put the two shots together they may not work in stereo any more. So it’s just careful planning. It’s a brave new world.
F&V: Do you have any advice for people entering this brave new world of stereo 3D?
BH: I keep telling people who are interested in doing 3D that they have to step away from the video monitor for a minute and remember how they see the world. The real world is much more analogous to making a 3D movie than all these techniques of using depth of field and crazy camera gags to draw your attention someplace. Because as soon as you step away from the filmmaking process, our world is in 3D and everything is in its place. You just have to honor that when you go into 3D.
Did you enjoy this article? Sign up to receive the StudioDaily Fix eletter containing the latest stories, including news, videos, interviews, reviews and more.
Leave a Reply