THE MOVING CAMERA
By Garrett Brown
In the second and concluding part of his series of articles for Zerb, Garrett Brown, Oscar-winning inventor of Steadicam and a whole range of highly innovative camera equipment, continues his exploration of just how and why the moving camera, when used both skillfully and appropriately, can add immeasurably to the dramatic impact of a shot or sequence.
Cameras everywhere are on the move. Some are being transported by inventions of mine and I’m content, although a few of my customers took so long to get on board that the early patents were nearly expired! After 18 years the ultra-conservative National Football League in the USA has not only approved the aerial, wireborne Skycam, they have embraced the gutsy, low-level, pursue-the-quarterback shots we dreamed of back when the prototype soared over my old high school field in 1983. (I did literally dream of these shots as a kid, except that I was personally doing the flying - kicking my legs through the viscous air and holding a camera to my eye!)
Most of my recent ‘cams’ - Moby-, Dive-, Go- and Fly- (if not yet the apocryphal Mole-cam!) - are out working and it’s fun to devise never-before-seen traveling shots when new sporting events invite us in. Ice hockey and basketball are experimenting with aerial moves and Flycam has flown fast and low over Formula One and NASCAR pit-rows. In the UK, Skycam has shot rugby for the Six Nations Cup and is poised to do the FA Cup and the Challenge Cup for the game we Americans persist in calling soccer. I still dream of hovering just above rodeo, figure skating, ballet and so-help-me curling (although the air seems thinner these days and my legs tire more quickly!)
One of the problems we face in live television, is how to transcend the sameness of shots, even spectacular shots, that are used over and over for coverage. Some set-ups are ‘one-trick ponies’ limited to a single path and angle. Others permit almost unlimited variation, constrained only by the realities of the sport; ‘Skycam, stay behind the line of scrimmage!’ Our goal, for any and all means of camera motion (including ones-and-zeros) is to deploy them reliably, sensibly, purposefully, artfully and to good effect.
‘Reliably’ and ‘sensibly’ are matters of prudence and experience. ‘Purposefully, artfully and to good effect’ (particularly the latter) are the considerations of this essay.
As stated in Zerb’s last issue:
Why is one moving-camera shot so much more moving than another?
The first installment offered several leads: from the simple 3D effect to kinetic phenomena such as the ‘appearing’ point, etc. The text of that first part and its video examples will remain up on the web, to be joined by the text and clips for this concluding segment. In Part II we will look into the visual and dramatic implications of various moving-camera choices. I’ll point you to some relevant movie sequences, ‘moving’ and otherwise, and one elegant little shot from the film Carlito’s Way in which a number of non-obvious manoeuvres had remarkable results.
Zerb Part II (Clip 1 etc) identifies the brief video clips that can be streamed on www.garrettcam.com. Other referenced sequences in Part II include time cues (__h,:__m:__s) within the respective films.
To begin with, let’s examine the one self-evident property of motion pictures - they illustrate small slices of time. They appear at a rate that seems continuous but the images don’t necessarily ‘move’ until some aspect of the frame begins to change. That motion may be of the subject itself, or of the ‘taking’ instrument, moving in any of three spatial dimensions (up/down, left/right, forward/back), in angular rotation (pan, tilt, roll), or in magnification (zoom in/out) or any combination thereof.
Each of the above changes of state has its own provocative qualities. But, physically moving the camera, with its manifest penetration of space, has the potential to be more informative and/or gripping than merely panning, tilting or zooming which might otherwise be accomplished on the rostrum camera with stills. Fortunately, in the live-action world, unless scrupulously centred on the nodal point of the lens, even panning and tilting exhibit spatial motions that are much more like the way we actually see - just try panning your head on the ‘nodal point’ of your eye.
Consider a person seen in the simple act of sitting down. A nearby camera might just tilt and our angle on the background will correspondingly alter. Or, we might boom down in sync with the sitter; but, as we demonstrate at Steadicam workshops, if we are too successful, if the lens height lowers evenly with the actor’s head, it creates an odd illusion - that the person has inexplicably shrunk or the set has risen up. However, an adroit combination of tilting and booming retains the sense of the original angle, particularly on the face, while making it clear that the actor has indeed sat down.
Rather than just panning as an actor walks by, we might also dolly out of the path and back in behind to more of an over-the-shoulder position, or add a parallel move for emphasis (countering in the reverse direction speeds up the cross while moving with the person steals energy, giving a more legato change of perspective); and a bit of booming in the middle avoids tilting up to the ceiling. These combinations require muscle and forethought with a dolly, but a good pedestal or Steadicam operator will instinctively make adjustments to affect the look and feel of a scene. At the very least, controlling foreground vs background can help keep your lights out of frame!
Regarding zooming: I came of age in the movie business operating Steadicam with prime lenses. On One From the Heart, we tried the Cooke 20-60mm on my BL III, but it had to be zoomed by my assistant, and the combination was mortally heavy. Finally, on La Traviata, Live from Paris, 2000 (also for Vittorio Storaro), I got control over my own lens and could move and zoom at will. In a tiny apartment on Isle St Louis, I had the joy of shooting the entire finale of Act IV in a single 23 minute uncut one-take Steadicam shot, as Violetta approached her death at the window on the stroke of Notre Dame’s midnight bell. I made hundreds of undetectable adjustments in focal length, watching the images unfold on the monitor, as moment, place, lens angle and motion came together time and again. Though perhaps routine for our brothers and sisters who operate pedestal-mounted TV cameras, the understated art of the move/zoom combination was a revelation.
Disembodied, ‘omniscient’ storytelling viewpoints like the above can not only cut from one place to another but may also move along to accompany the action in whatever way best tells the story. These ‘objective’ shots tend to travel smoothly and are unconstrained as to course, speed and position. POVs, however, are ‘subjective’, as if from some entity’s point-of-view, with angular or spatial irregularities that may suggest the rising and falling and darting of its ‘eyes’ as it creeps, leaps or flies.
Easily ‘overacted’, the POV effect thrives on restraint. The scariest moment in Silence of the Lambs is the night-vision-goggled fiend’s last infinitesimal move, closer still to the blindly-reaching Clarice (1h:47m:31s).
When in doubt …don’t! Save it, keep it in reserve. As my classical guitar teeacher once inelegantly put it: ‘Don’t piss on every fireplug’ (or in your more civilised parlance: ‘Don’t seize every opportunity to impress’). In a widescreen POV, for example, you can give subjects some freedom to roam within the frame and not slavishly pan to keep them in the middle - like the ‘eye-motion’ rather than the ‘head-turning’ component of your entity’s point-of-view - and reserve the extra emphasis of centring the action for when it really counts.
My favourite examples in the last few years, like the above Silence of the Lambs sequence and the witty POV in Terminator II (6m:35s) when naked Governor Schwartzenegger enters the bar, have tended to be accomplished with Steadicam. I generally dislike the handheld camera for this purpose - its angular shakes are a bit too overt. Our in-built human stabiliser gives our own vision a smoother look, more realistically approximated by a platform that is both stable and free to move about. (Of course I point this out as a public service, without any taint of commercial interest!)
My colleagues and I once had a go at duplicating the holy grail of POV effects - the way it ‘looks’ when we dart our eyes and turn our heads quickly from one object of interest to another. When we do this what do we actually see? How might it be reproduced with a camera? For a Molson Beer commercial I directed, Larry McConkey and I modified my Steadicam by minimising its inertia and increasing the size of the handgrip so it could be whip-panned as quickly as possible. We hooked up a video camera and Larry began practising.
An ultra-rapid pan is an energising transition device in the continuity of movie and TV scenes, though executing good ones and landing them precisely with conventional gear has always been tricky. ‘Whip-pans’ had been a spectacular addition to the Steadicam repertoire and Larry could really smoke ’em with our modified rig, although the end frame was correspondingly harder to control. Our tests did sort of resemble the ‘swish’ POV of one’s eyes but they still weren’t fast enough. The solutions to both these problems have proved quite useful ever since.
Speeding them up was easy. We just cut out a few alternate frames near the centre of the pan. We also found that different takes of the entire scene could be seamlessly tied together at the midpoint of these darting ‘looks’; and as long as the mating whips actually occurred in the same place within the set, the cuts were undetectable (cheating the locale or using generic ‘swish’ frames is rarely convincing). Similar mating cuts can be hidden within speed-ramping effects and snap-zooming shots, just as they have been more or less concealed by dollying ‘wipes’ since the early days of film. These midpoint cuts permit long sequences to be invisibly broken up and executed piecemeal. My cinematic eyeball darted its beer-loving ‘gaze’ here and there four times during the course of our ‘uncut’ shot, and the cumulative impression is humorous and does resemble what we do with our eyes! (clip 1)
Shakes, Turbulence & Wide Angles
Another visual phenomenon is the effect of rough camera motion on one’s movie-going perceptions. The shaky handheld look, for example, may now proudly signify ‘lack of artifice’ in low-budget or ‘Dogma’ style movies. Formerly a nod to ‘New Wave’ features and the shoulder-held documentary tradition, it lately also has suggested the ‘camcorder mini-DV’ look.
Extreme angular vibration is a hoary convention that suggests out-of-control danger - to the unsteady viewer or from the unsteady environment - such as the violent shaking that accompanied cheesy earthquake scenes and horse and buggy chases in the Saturday matinees of my youth. The resulting blurriness also handily distracted viewers from spotting phony aspects of the set or the action. The effect is less fashionable these days, reserved for severe hand-to-hand combat etc, and digital artists can now shake their virtual cameras with just a couple of keystrokes - especially, perhaps, when they too have something to hide!
However, when the camera angle remains unchanged but the camera dodges around in space, a different and very useful illusion can result. On Twilight Zone: The Movie we found that if I kept the camera pointing in the same direction while booming energetically up and down and/or lurching side-to-side with the Steadicam, it produced the specific impression that the plane set (which was immobile) was plunging and side-slipping due to extreme turbulence The effect was brilliant.
Regrettably, foolishly, I tried to improve on it by asking the extras to rise up against their seat belts when they peripherally saw me lift up the camera and to sink down into their seats when I boomed down. It’s in the movie and it looked perfectly convincing until the leaping, lurching lens passed by a little girl actress who hammed it up totally destroying the illusion. (I hope by now she has experienced some real turbulence!)
What else? My grab bag of visual considerations includes a way to move extreme wide-angle lenses so they don’t give themselves away by ‘keystoning’ the verticals. On The Shining, Stanley Kubrick found that if I rigorously kept my camera level fore and aft he could get away with a 9.8mm Kinoptic lens in the Maze sequences. I installed a bubble level lengthwise on the camera and kept the lens down at mid-height - lower than we are used to - and merely boomed up a bit to control the headroom when Shelley and Danny came close. The Maze looked amazing, twice its real size, yet it wasn’t obvious that we were using an ultra-wide lens. By the way, it helps to take classic ‘racing’ turns, seriously cutting the corners, and thus panning as gradually as possible.
So here’s the punchline: I believe that camera operating can have an impact on the dramatic arc of a scene of the same order as music, lighting or art direction I don’t mean just pointing the thing but the whole complex job - the indispensable coordination of timing, camera motion, mechanical execution and action; the camera’s putative ‘presence’ and behaviour; the acceleration curves, staccato or legato; the intersecting visible and emotional trajectories of persons, places, things; the ‘punctuation’ of events … and that’s just for a locked-off tripod shot! (Only kidding!) Oddly enough, lazy, pedestrian, clueless operating may be harder to spot because it’s so uninvolving that we can’t be bothered to pay attention.
The differences between a ‘nothing’ move and a great move suggest to me that the intrinsics of great camera operating are analogous to those of ‘violin operating’ - the pitch-perfect moments, the mastery of tempo, the modulations, rests and full stops, the inflections and dynamics. Like a virtuoso performance compared with a student recital, the impression is cumulative - as when an instrumentalist varies the ‘colour’ of the sound or imposes ‘rubato’ (robbing time from one phrase to add to another) for emphasis - and we, the audience, who have moved all our lives and are savants of visual motion, will surely get it!
When, at the end of Steadicam workshops, we instructors have a go at the ‘Final Test’ shot, inexpert versions of which we have watched all day , it’s amazing to feel the emotional and dramatic voltage ramp up due to the sudden novelty of idea-directed framing and the Zen of starting and stopping, closing in and pulling away in telepathic counterpoise with the actors.
Having taught more than a thousand Steadicam operators, young and old, I am extremely aware of the challenges presented to the novice. As my friend, the madly eloquent French operator Jean-Marc Bringuier put it in his Gallic 1989 polemic for Cahiers du Cinema (his translation):
Every serious camera operator will agree that fine-tuning his manual adaption to the traditional camera supports, fluid or gear heads, is but a minute part of his task. As a demanding front-row spectator, he is concerned, above all, about making his visual and auditive attention even acuter and more responsive to every nuance of comedy…Using a carried-and stabilized system such as Steadicam implies a thorough and frustrating re-learning of gestures, shooting postures and ways of monitoring the shot. But it does also entail, as a dreadful consolation, that novices quickly reach a point where they can start turning junk into tinsel with a vengeance. This sprightly regression towards gaudiness swiftly sweeps away all previous subtleties …like a dolly going on a spree, after having slipped away from its tracks.
I didn’t much enjoy reading this at the time, but he’s right. Like the trombone, which I briefly studied, the Steadicam at first seems maddeningly imprecise - no frets, no keys, no repeatable marks or frames - which makes for some astonishingly random performances. It is only as we become intimate with the Newtonian verities of the beast that we learn to achieve specific outcomes by sheer force of will. Those of my early shots that succeeded, like that first stroll off a crane into a mob of 900 extras on Bound for Glory required fierce concentration. And those that went wrong, including the wobbly finale of the famous ‘Rocky up the Stairs’ shot, reveal distractions of some sort (our beloved DOP Ralf Bode falling a bit behind with the two car batteries needed to run my damaged prototype!)
Over the years I have collected moving-camera sequences, admirable and otherwise, to show at workshops, and both categories are instructive. In light of our preceding discussions here are a few examples that are easy to obtain on tape or DVD and I have included some time cues, in case you care to look them up. (Also, as promised, several can be streamed as Quicktime video clips on www.garrettcam.com).
One of my favourite illustrations of the simple 3D effect is Greg MacGillivray’s opening helicopter sequence for The Shining. Try hitting the pause and play buttons repeatedly as he flies over the aspens (1m:40s).
As to the ‘kinetics’ referred to in Part I, my very first day’s Steadicam work on the same picture (3m:2s) is a good workaday example. The lens sticks with Jack Nicholson and sidles with him through the set in a most satisfying way (so what if it’s take 50!) - as likewise Liz Ziegler’s first shot (0m:54s) on Kubrick’s Eyes Wide Shut, made at the beginning of her year and a half of work for Stanley on the one film he never got to complete.
Sometimes a good Steadicam sequence inspires years of earnest homage. Nicola Pecorini’s shots for John Seale on American President were concise and effective, but any direct comparison with TV’s The West Wing must reckon with the realities of episodic television and the temptation to gulp down numerous pages of dialogue at one go.
One of my all-time favourite action sequences is the subway and train station chase that concludes Carlito’s Way (2h:4m:16s). Watch Al Pacino’s Carlito try to evade, outrun or outwit his pursuers in the bowels of Grand Central Station in order to join his girlfriend and escape to Florida on the train. Director Brian De Palma, DOP Steve Burum and operator Larry McConkey used Steadicam as an ensemble instrument, intercut with crane, dolly and tripod, and the result is masterful and infinitely greater than the sum of its parts - heart-pounding, relentless, pell-mell with witty and violent change-ups, it’s one of the best chase scenes ever. Six minutes in, note the sensational uncut 2-minute Steadicam shot (2h:10m:12s), see it morph in and out of Carlito’s POV (2h:11m:45s), and then give us something quite astonishingly original in McConkey’s surreal aerial flight down the escalator (2h:12m:2s).
Another marvelous little anamorphic shot of Larry’s in the same movie (1h:44m:14s) deserves to be seen (clip 2) to appreciate its perfect track and off-centre framing (and ‘disappearing point’) and the brilliant blocking. Watch Penelope Ann Miller’s Gail press the right frame edge, while Carlito holds back and then eases himself into an ‘over’. Larry repeatedly ups the ante by imperceptibly edging in, and his sudden full stop makes Gail’s line, ‘I’m late!’ into the emotional high point of the scene. It’s fabulous!
On the other hand, though I greatly admire Kenneth Branagh, there are a number of non-admirable uses of Steadicam in his Mary Shelley’s Frankenstein. In light of the notions presented in these articles, check out Branagh’s ‘walk & talk’ with Tom Hulce (23m:20s to 24m:30s) or (clip 3).
Finally, if I’m ever tempted to believe that Steadicam is indispensable, please let me recall that most terrible night on Scorsese’s Bringing Out the Dead, when we learned that the after-market Panavision mount on my rental Moviecam had an undersized aligning pin. The anamorphic lens had twisted in the mount and caused parallelogram distortion in my previous shot. I had been poised to undertake a sensational Steadicam sequence but now DOP Robert Richardson prepared to shoot it conventionally on a dolly. As I sat on the sidelines, eating my liver with frustration, Richardson (a superb operator), his focus puller Gregor Tavenner, and dolly grip Patrick McGrath rehearsed six or eight times and then simply made it happen. Nicolas Cage raced down a hospital corridor preceded by the hurtling dolly and dashed into the ER with a dying baby in his arms. Richardson hung precariously on to the jib and worked his fluid head as the it rose and fell, and it was the quintessence of what the artisans of ER and West Wing have always been striving for. The sequence (clip 4) begins 59 minutes 47 seconds into the film and among a number of perfect moves, I direct your attention to the moment, just after the camera whips over to a flat-lined heart monitor, when it slowly, slowly pans back - past the dismayed doctors to the retreating, stunned figure of Nick Cage. It is the emotional crux of the scene and it’s the shot design and the operating that carries us there as the grim tension decays into something infinitely more bleak. I show this shot at workshops to dispel any notion that Steadicam confers some kind of unique magic upon us. It is among the handiest of camera-moving tools, but using any of them well has more to do with one’s moviemaking sensibilities. Period.
In this age of video assist, with gaggles of second-guessing executive producers huddled in ‘video villages’, the skill, experience, confidence and discretion of the old-school operators are needed more than ever behind camera. Raise your glass to the late Mike Roberts, the gifted (non-Steadicam) UK operator whose work on fifty movies, from A Man for all Seasons to Chocolat continues to impress. Likewise Peter MacDonald, with whom I had a friendly rivalry on Yentl. He and I were, in effect, competing for screen time - his mastery of the conventional filmmaking hardware vs my Steadicam chops. On one priceless occasion we both got to shoot the same scene and directly compare the results. Peter laid his rails, and made his elaborate crane moves. Then all was struck and I did it all over again on foot; laddering up onto wagons and stumbling off onto apple boxes, high-stepping my way through the weeds. Director Barbra Streisand printed both versions for uproarious comparison in the screening room and intercut between us in the movie. It was magic.
Roberts and MacDonald came along in the footsteps of Brit camera operators like Alec Mills and Chic Waterson - stalwarts of the ‘English system’, in which the operator’s influence on shot design is large. They did their stuff without being ‘supervised’ on video, and used gear that was more cumbersome than our present-day Technocranes and Steadicams.
So here’s to you Alec and Chic and Peter and Mike and all our colleagues around the world behind film and TV (and mini-DV!) cameras - what really counts has always been and will remain the same: To watch closely, to listen to the music, to comprehend the heart of the shot and the soul of the story. The tools are analogous, the results are equally meaningful - the momentary angle of the lens, the quality of the move, the manner of starting and stopping, and always the rigour of when and why and how – whether for Beowulf or Baywatch or your kid’s birthday.
The handheld camera is an instrument; the Steadicam and the crane and the dolly are instruments; capable of infinite discretion of choice; movable certainly, but not necessarily moving…
That part is up to us.
As well as inventing Steadicam and many other moving camera devices, Garrett Brown has operated Steadicam on a long list of Hollywood movies. For credits and biographical information, details of his unique range of camera equipment, or to watch the clips referenced in this article (available from 7 March) visit www.garrettcam.com