Figure 1. The start of a dolly forward-zoom back shot. Camera with telephoto lens films two persons standing side by side. A telephoto lens captures an image along a narrow angle, so in order to fit both actors into the frame the camera must be placed a moderate distance away. Even so, because the narrow angle the image subtends a narrow slice of the background behind the actors.

Figure 2. The end of the shot. The lens has zoomed from telephoto (narrow angle) to wide angle. Because of the wide angle the camera must move closer to the actors to keep their images filling the frame. But look at the background! In spite of the camera’s forward movement, the image subtends a much broader portion of the background.

Photos A, B, and C illustrate the same process.

Photo A. CSICOP Senior Research Fellow Joe Nickell stands in profile, pointing. His image fills the frame. This image was made with a telephoto lens (135mm lens on 35mm camera) from a substantial distance away. Note how little background is included, which is why Skeptical Briefs production editor Tom Genoni (out of focus in the background) subtends about a quarter of Joe’s apparent height.

Photo B. Made with a 50mm ("normal”) lens. I moved substantially closer to keep Joe the same relative size in the frame. We see a lot more background; Tom has shrunk appreciably.

Photo C. Made with a 24mm (wide angle) lens. I’m only a few feet from Joe, and the entire Center for Inquiry stands tiny in the background. Tom is almost invisible.

Imagine a motion picture sequence that moved through this range. The apparent separation between Joe and background objects would increase dramatically. The emotional subtext would suggest Joe being wrenched from his surroundings, or perhaps his environment fleeing from him. Powerful stuff.

Photos D, E, and F show what happens when one zooms in while keeping the background roughly the same size.

Photo D. Shot with telephoto lens from about 75’ away. Library windows fill the image from side to side. CSICOP staffers Marsha Carlin and Etienne C. R'os seem to occupy the same plane though Marsha stands about 12’ in front of Etienne. (See where their feet are!)

Photo E. Shot with a normal lens. Marsha and Etienne have not moved, yet their apparent separation has ballooned.

Photo F. Shot with wide angle lens. I’m only about 2’ from Marsha, too close to hold her in focus! She and Etienne seem to be in different zip codes. Both also seem much more distant from the building.

Last issue, we examined the once-difficult, now-routine motion picture shot in which the camera dollies forward while zooming out at a rate which maintains the foreground subject at a constant size. The eerie result: While the foreground character remains stationary, the background flees outward in all directions. It’s a great way to express sudden isolation or dramatize a character’s response to some shocking revelation. No sooner did I finish writing the last installment than I saw Ron Howard’s Apollo 13. That film uses the dolly forward-zoom out technique for a brief reaction shot of flight director Gene Kranz (Ed Harris) at the moment when the astronauts report their emergency. It’s sound movie-making — and one more indication that this formerly-exotic device has become an accepted part of film grammar.

But what makes it work? Consider the effect of lens length and camera-to-subject distance — in a word, of perspective — on the way a shot “feels.”

By selecting lens length and camera distance wisely, movie and TV directors can control the emotional resonance of their shots — creating subtle impressions of camaraderie or loneliness, enmeshing individuals in their environment, thrusting them into savage isolation, or placing a romantic couple in a “zone of their own” set off from their surroundings. It’s one of the strongest ways to influence audience response to an image, yet few suspect anything — until a director draws attention to the process by means of a bracing dolly in-zoom out shot.

Effects technology has made countless strides since Star Wars launched the revolution in 1977. Still, one of the major reasons fanciful creatures and objects look better in today’s productions is that effects artists have largely solved the problem of injecting suitable blur into their work. In this column, we'll see how.

For Star Wars, effects supervisor John Dykstra developed the first practical motion-control camera system for dynamic photography of spaceship models. Highly repeatable stepper motors drove a camera boom that swept over almost-stationary models. Multiple passes of the same shot could be made in perfect alignment: one to capture the model itself, one for its on-board lighting, one to obtain a perfect silhouette for later use in “matting” the model into a background image, and so on. With early motion- control systems, a two-second model ship fly-by might take hours to shoot. Camera speed was slowed down proportionately; for each frame, the shutter might open for several minutes while the camera crept past the model. During each frame the camera moved approximately as far relative to the model as it would have in a “real” shot. Near-perfect blur was automatic. The result was a more realistic impression of fast and violent motion than had ever been achieved before on the screen. In Spielberg’s Close Encounters of the Third Kind (1977), effects designer Douglas Trumbull used similar technology to produce realistically-blurred multipass images of glowing UFOs.

By the time Spielberg made his ill-advised John Belushi comedy 1941 (1979), technicians at Industrial Light and Magic (the Marin County, California, effects factory that grew out of Star Wars) were experimenting with a technique called “go-motion.” As the name implies, go-motion was a direct attempt to address the blur problems inherent in stop-motion photography of miniatures. Additional computer-controlled stepper motors were attached not to the motion control camera boom, but to the miniatures themselves. Instead of being posed by animators between frames and photographed at rest, go-motion miniatures would move, repeatably, at microscopic speed. In 1941 go-motion contributed to a few long shots of a Japanese submarine on the surface. Spielberg wanted to do camera moves over the sub and matte it into background imagery of the Pacific Ocean even though crewmen were visible on the sub’s deck. This would have created an insuperable compositing challenge if go-motion had not made the little crewmen on the model sub move exactly the same way in pass after pass.

Go-motion came into its own in Spielberg’s E.T., the Extra-Terrestrial (1982). Remember that film’s signature image: E.T. and his young human friend crossing before the moon on a flying bicycle? Go-motion motors repeatably rotated the bike’s spoked wheels, making possible complex composite shots where the smooth, realistic wheel movements “sold” the effect.

When ILM tackled Return of the Jedi (1983), the finale of the original Star Wars trilogy, go-motion technology wasn’t ready for the challenges it posed. (Jedi featured scores of composite shots with bright backgrounds: forests, deserts, smoky rooms with lots of backlight-far harder to composite than shots with dark outer-space or night-sky backgrounds.) To produce the Rancor, a fifteen-foot lizardlike biped that menaced Luke Skywalker in Jabba the Hutt’s lair, effects supervisor Dennis Muren rejected both stop-motion and go-motion. Instead, animators Phil Tippett and Tom St. Amand used puppetry and concealed rods and wires to manipulate the miniature Rancor, which was shot “live”-that is, actually moving-with slow-motion photography.

James Cameron’s The Terminator (1984) was the last major film to offer an old-fashioned, jerky stop-motion character. The Terminator robot (supposedly Arnold Schwarzenegger’s endoskeleton) was realized with life- sized puppets wherever possible. For certain long shots, there was no way to avoid stop-motion (and no money for go-motion). Animator Peter Kleinow used a Vaseline-smeared glass plate between the lens and the model to suggest blur, but it didn’t work.

Back on the high-tech front, The Golden Child (1986) gave ILM a chance to try out a new real-time motion control recorder. For a sequence of Eddie Murphy battling a man-sized demon, director Michael Ritchie shot live-action footage of Murphy fighting a non-existent opponent. The camera moved freely; in some shots it was hand-held. Effects technicians used field recordings of all that movement to apply precisely matching moves to their go-motion footage of a miniature demon. The shots were amazingly good, especially considering that Ritchie jerked his live camera more enthusiastically than the ILM gang originally had in mind.

Robocop (1987) showed that a gifted animator could get good results even with plain old stop-motion. Shots of the ED-209 “enforcement droid” were done stop-motion in front of rear-projected backgrounds-just the way Ray Harryhausen did films like Jason and the Argonauts (1963), whose failings I discussed in my previous column. Tippett, by then Hollywood’s master stop-motion artist, added convincing blur in a refreshingly low-tech way. While exposing each frame, “we introduced blurs basically just by wiggling the puppets,” he told Cinefex. It was the last sustained used of stop- motion in a major Hollywood picture, and it worked remarkably well.

When Spielberg started planning Jurassic Park (1993), effects artists planned to execute long shots of the T-rex, velociraptor, and other dinosaurs using go-motion miniatures. Advances in computer graphic (CG) animation persuaded the makers to abandon go-motion in midstream. A huge ILM crew under supervisor Dennis Muren realized the full-body dinosaurs as perfectly- realized three-dimensional computer constructs. (Continuing the vocabulary of “stop-motion” and “go-motion,” they called the new method “full- motion.”)

Mathematically-exact blurs were incorporated right into the images as the computers rendered them; as every living human knows, the results were perfect. Phil Tippett had been hired to direct the go-motion work; instead he shared with the CG artists his deep understanding of how to make artificial creatures “perform,” and helped develop a number of “waldoes” (wearable hand or body rigs that let animators feed motions to their computers in a more lifelike way than by typing start points and end points into a computer console).

Another ILM crew under supervisor Mark Dipp used Jurassic Park technology to create more fanciful CG dinosaurs for The Flintstones (1994). One “must-have” scene replicated the familiar cartoon gag in which Dino, Fred Flintstone’s purple pet dinosaur, drags Barney (Rick Moranis) across Fred’s living room at the end of his leash. The background shot was a blurry pan shot that followed Barney across the soundstage living-room set. For the computer, applying proper blur to Dino’s movement plus matching the blurs introduced by the background camera’s movement was a piece of cake; stills from this sequence are amazing in the realism and correctness of the blur they displayed.

Where do we go next? High-end CG imagery is migrating onto simpler and less costly computers. Instead of the high-end Silicon Graphics workstations used in Jurassic Park and The Flintstones, several shots of the Enterprise-D starship in Star Trek: Generations (1995) were created entirely as 3-D CG constructs. They were rendered, blur and all, on ordinary Apple Power Macintoshes using off-the-shelf software by ElectricImage, Inc. Rendering at motion-picture resolution took an average of just six minutes per frame.

Next time you go to the movies, don’t expect jerky, failed stop-motion shots to tell you how the shots were done. Hollywood’s effects artisans have long understood that convincing motion requires not only changes of position in successive frames, but appropriate blur as well. And the problems of creating it have been conclusively solved. Heck, by the time this column sees print you may be able to do it on your own desktop.

Effects pioneer Willis O'Brien made the giant-creature shots in King Kong using a technique called stop-motion animation. Stop-motion animators pose flexible model creatures (apes, dinosaurs, what have you) in miniature settings and expose a single frame of film. Then they move the creatures into the positions they would have assumed 1/24-second later if actually in motion, expose another frame, and begin the dreary round again. It’s laborious, but for decades it was the only way to bring truly outrageous objects “to life” in the cinema.

The thing to remember about traditional stop-motion is that none of the “moving” subjects are actually in motion when they are photographed. But an illusion of motion arises when the human eye views these still pictures, twenty-four of them each second, on a motion picture screen. (The numbers differ for video, but the same principles apply.)

Stop-motion is far from perfect. To see why, we must consider how motion picture images are created—and how the brain interprets them to produce the illusion of motion. A motion picture camera operating at sound speed exposes twenty-four frames per second (FPS). It’s a purely mechanical process— advance the film a frame, stop it, open the shutter, close the shutter, advance to the next frame—so less than half of the time between frames is available for exposing the film. The actual exposure time for a frame of motion picture film is about 1/60- second. If you've used a 35mm still camera that lets you set your own shutter speeds, you know 1/60-second isn’t very fast. If you photographed a basketball game at that shutter speed, you'd get blurry images of the players. To get stills that freeze the action, you'd want a faster shutter speed — say, 1/500-second.

Why do moving objects blur with exposures in the 1/60- second range? Imagine photographing a passing car. At 60 mph, the car moves 88 feet per second, almost a foot and a half during the 1/60-second that your shutter is open. How will the car look when the prints come back? It will blur, of course, reflecting the fact that the car was not in the same place throughout the exposure. Blur severity varies not with an object’s absolute speed, but with how much of the image area it crosses during an exposure. Set your shutter at 1/60-second, stand 200 feet from a superhighway, and snap a picture. The cars will only blur a little. Stand on the shoulder and snap that same traffic, and you won’t be able to tell Fords from Toyotas.

When it comes to blur, Hollywood movie cameras work just like your still camera at 1/60-second. They just cost more. Pop your favorite movie in the VCR. Freeze a single frame of an action scene. You'll see blurring you never imagined was there. But your brain notices it. More, it expects moving objects to be blurry.

That brings us to the question of how the brain interprets the projected motion picture image. Everyone knows the basic principle: Successive still images are flashed on a screen, and a phenomenon called “persistence of vision” keeps us from seeing the intervals of darkness between frames. We view the succession of stills as a continuous image. When an object changes position from frame to frame, we perceive that the object is in motion. But this isn’t the only cue that can fool the visual system into perceiving movement. 3-D movies exploited stereoscopic vision to create vivid impressions of movement. Conventional movies don’t take advantage of stereopsis. But they do take advantage of other assumptions the brain seems to make about moving objects. One such assumption is that the image of a fast-moving object will be degraded as a consequence of its movement. In other words, if an object is moving quickly enough across the visual field, the brain expects it to blur.

Keeping that in mind, we can reconstruct why the pilot’s- eye view shot in King Kong looked so good, and why the smash- the-airplane shot looked so bad. Since subjects in classical stop- motion do not move during exposures, they do not blur. For the pilot’s-eye view shot, Willis O'Brien rigged a stop-motion camera to roll down a track over a huge New York skyline diorama toward a model Empire State Building. Though the camera seemed to move at 150 mph, since the objects in motion relative to the camera (the skyline) were distant, nothing moved very far across the image area between any two frames. If the shot had been staged for real, you wouldn’t expect much blur. So the fact that the stop-motion sequence had no blur did not detract from the illusion of motion it created.

The smash-the-airplane shot was a fairly close, static shot. The airplane sped across the screen; Kong’s arm lashed out and struck it. With each frame, they crossed large fractions of the image area—normally a recipe for severe blurring. But stop- motion can’t blur! When viewing this scene, we experience conflict between two modes of visual interpretation. Objects change positions drastically from frame to frame, which tells the brain that they are moving quickly. But the absence of blur tells the brain that everything is stationary. Result: interpretive conflict. The illusion of motion is compromised.

Stop-motion shots that don’t have the blurring they need give viewers a “strobing” sensation. Objects jerk-jerk-jerk like dancers under an old disco strobe light. Three famous scenes created by stop-motion master Ray Harryhausen in the 1950s and 1960s exemplify the problem. Next time you find Jason and the Argonauts on late-night cable, watch how the swordfighting skeletons “strobe,” especially in close shots. Check out the jerky movements of the giant crab in The Mysterious Island. In One Million Years B.C., a stop-motion pterodactyl carries off Raquel Welch. Its wings flap in and out of frame at high speed—but with no blur. Even to untrained eyes, it looks profoundly wrong.

In the next installment, we'll see how moviemakers since the time of Star Wars have applied high technology—and sometimes, startlingly low technology—to inject blur into animated footage. Understanding how and why Hollywood professionals use blur to make their illusions more effective can help us all understand the myriad ways the eye can be fooled.

What’s going on here? We're catching our eyes in the act of second-guessing the color of everything we see. Even skeptics tend to think of vision as a simple, linear process. Researchers probing the details of vision say otherwise. Francis Crick, codiscoverer of the structure of DNA and a CSICOP fellow, devoted a book to summarizing vision research and its implications for understanding human consciousness (The Astonishing Hypothesis, Scribner’s, 1994). Visual experience is profoundly synthetic. Raw data from the retinas undergo extensive processing and interpretation, starting in the retina itself. Structures throughout the brain join in shaping what we seem to see so naturally through our eyes. But seeing is not believing.

The red-squares illusion illustrates the Land effect (discovered by Edwin Land, who went on to invent Polaroid photography). Here’s the problem: We never see the “actual colors” of objects. All we get to work with are the wavelengths of light they reflect. that reflected light can be affected by the object’s actual color or by the color of the light under which we view it. Or both. White paper under red light looks red. Red paper under white light looks red too. Yet we can usually tell them apart. How?

Apparently, still further up the visual-processing ladder, color information from one part of the visual field (the paper) is compared with that from the background. The brain tries to subtract the effect of colored illumination to restore the “actual” color of objects. That is the Land effect, and it’s why a white jacket still looks white instead of orange under the setting sun. Like most of the brain’s subterfuges, the Land effect can be fooled. Viewing those two red squares on different-colored backgrounds tricks the brain into thinking that the red square on the blue background actually lies under blue light. So it “compensates” by changing the square’s perceived color.

What’s remarkable is not that the brain’s color compensating apparatus can be fooled, but how well it usually works. We live our lives under many different colors of “white light” and hardly ever notice. Ever take snapshots indoors without flash? The photos come out with a strong orange cast. Were you about to say an “unnatural” orange cast? Bite your tongue! That ruddy tint is altogether natural. The light from a household table lamp really is that much redder than sunlight.

Professional photographers measure the color of “white” light in degrees Kelvin (degrees Celsius above absolute zero). “Color temperature” corresponds to the color of light that would be radiated by an ideal black body heated to that temperature. Lower temperature means redder; higher temperature means bluer. As shown in Figure 1, household incandescent lighting has a relatively low color temperature-about 3,000° K. (Only open flames-candles, campfires-are lower.) By contrast, direct sunlight has a color temperature of about 5,400° K., much bluer.Photographic film can’t compensate for the color of ambient light like as your brain does. Your film reproduces color most accurately at 5,400° K, the color of direct sunshine (and photoflash units). Your available-light indoor photo is taken at 3,000° K. You don’t see the color difference, but your film does. Result: orange snapshots.

Real life is a symphony of color temperatures most of us never see (Figure 1 again). Incandescents used in television and movie studios are bluer than household lighting, about 3,400° K. (Professional motion-picture films are balanced for this light; in low-budget movies, the view from the livingroom window often has a blue cast. Producers with more money put orange gels over the windows.) Open shade is about 400° K bluer than sunlight. Some of the high-intensity lights now used in Hollywood hit 6,200° K. They require orange gels to match sunlight! Outdoors at night, these units are often used without gels for maximum light output. That’s one reason that artificially lighted night exteriors in movies like Terminator 2 look so blue.

Since the 1970s, Hollywood has delighted in teasing us with the colors of white. Letting the colors shift suggests gritty realism: Remember the orange interiors of the Godfather films? Consider the way police station interiors are often lighted in cop movies. Most of the light looks white (3,400° K). The windows go blue (5400° K). And the little fluorescent lamps on everybody’s desk look green. Green? Color temperature isn’t the only variable in “white” light. Most fluorescents have a green spike in their color spectrum: They emit a disproportionate share of their light at green wavelengths. Untrained humans never see that extra green, but film does.

Why don’t home videos show the colors of white so strongly? Camcorders compensate for the color of ambient light much as the brain does. As for professional videographers, they point their lenses at a white card and push a "white balance” button. It’s their way of telling the camera, “Hey dummy, this is white.” The camera then adjusts its red, green, and blue gain until the card reads truly white, compensating for color temperature and color spectrum in one operation.

What are the lessons of all this? First, if you shoot snapshots indoors without flash, stop waiting for the magic day when your pictures will stop coming out orange. They never will. Second, the next time you’re tempted to accept the evidence of “your own eyes” without additional corroboration, remember about the color of white. It isn’t just your eyes that see, it’s your brain — and some of the tricks your brain performs to “improve” your visual experience can distort it as well.