What Can We Learn From 2023’s Oscars Cinematography?
We often talk more about pixels and bit rates than we do about lighting, blocking and framing. But now’s the time to shelve the less technology-centric aspects of cinematography, and focus instead on this year’s nominees in the Academy Award for Best Cinematography category—and what we can learn from them for our own future projects.
The techniques of Oscar nominees might seem inaccessible, but if we look at them the right way, we discover the best thing about craft skills: they don’t necessarily require that your camera cart contains the entire contents of an ARRI showroom to be relevant. So, let’s stop staring at numbers and take a good look at the sort of images which are in the running for a little gold statue.
(And when you’re finished here, make sure you check out our epic workflow guide for all the films in the 2023 Best Picture category.)
All Quiet on the Western Front
Cinematography by James Friend, ASC, BSC
There’s probably no way to put World War 1 on screen that can’t be described as sobering, and All Quiet… is no exception. The film is unusual in that it looks at the war from a German perspective, with Friend’s Oscar-winning cinematography depicting both the literal and figurative chill embedded in the Erich Maria Remarque novel.
Day scenes were shot on ARRI’s large-format cameras, with the Sony VENICE used for its higher sensitivity at night. The choice of a large sensor is occasionally clear in the shallow depth of field of even quite wide-angle shots, especially as very little of the film appears to have been shot on anything longer than about a 50mm lens. All Quiet… features some deliberate framing and long takes, evidently with active camera stabilization.
None of that’s unusual these days. However, there’s something in this film that would be challenging to anyone—whether they’re shooting on an Alexa LF or an iPhone—and that’s its huge daytime exteriors, on a vast location (visible on Google Maps). These are often shot beneath overcast skies with little color contrast to lend separation to anything. There are ways to control this and ways to accept it and use it to our advantage, and All Quiet… looks to have done both and more.
Some productions bring out large lights in situations like this, adding shape to what’s otherwise the world’s largest soft toplight. For example, one of the biggest single source lights around, a 200-kilowatt Softsun, was used for the lunar surface scenes in First Man to simulate the infinite fall of sunlight over the vast, empty moonscape.
Most movies don’t need quite that much. But even so, lighting bigger areas means exponential increases in power. It quickly becomes impractical, even for the high end. Similar problems crop up for many productions even at night, where there may be a desire to see something of, say, a moonlit woodland. Even in the dark, the biggest lights must fall off to black eventually, and that falloff easily becomes visible.
All Quiet… doesn’t fight the inevitable, but it does use the sun to create a separating backlight. Here, the importance of scheduling is clear. In extreme cases, big frames of diffusion can turn unwanted sunlight into cloud, or black fabrics can create darkness—negative fill—where there’s a need to separate things by brightness. But the sun goes where it goes.
The best part of fabric is that it doesn’t need to be plugged in. It does, however, suffer the same scale problems as big, big lights. Flags and diffusion become less practical as the shot scales up—what works in closeup or single will fail in a drone pass over no man’s land. Smoke and haze is also widely used. It’s a particularly powerful way to create depth over large areas that doesn’t have to cost a fortune.
“Flags and diffusion become less practical as the shot scales up.”
One thing All Quiet… generally doesn’t use to create separation in its overcast scenes is overt color contrast. And that’s a factor not just of camerawork but also production design. One of the most powerful things productions at any scale can do is exercise some control over the color palette.
A film like this could plausibly have looked warmer, as was done for the World War 1 film Journey’s End. There clearly wasn’t a creative reason to do that, but All Quiet… was never going to be an explosion of party colors either way.
War movies inevitably put everyone in uniforms which are specifically designed to color-coordinate with the surrounding world. Production designer Christian M. Goldbeck and costume designer Lisy Christl realistically had few options other than browns and greens for much of the film. That’s a blessing and a curse. It makes for a consistent look, but, handled poorly, turns the frame into a sea of unappealing mush.
Compare 1917, or the 1980s Aces High, or even Saving Private Ryan, depicting the following war to equally greenish-brownish effect. Even Peter Jackson’s They Shall Not Grow Old, built from colorized monochrome footage, ends up with a similar palette. There are even color parallels with Aliens, which aped Vietnam war movies.
Whether this is what World War 1 looked like, or whether it’s what modern cinema audiences expect it to look like, is another matter.
So the production creates contrast with lighting and framing, which is complicated, and by shooting at sunrise, which makes everyone get up early. Of course, there are rich and involving interiors, bunkers, night scenes, and a million situations not lit by overcast. All Quiet… is worth dissecting for all that and more. Really good day exteriors, though, often demand a lot of effort, and can look easier than they are.
Bardo, False Chronicle of a Handful of Truths
Cinematography by Darius Khondji, ASC, AFC
Discussing the cinematography of any film ideally means forming a complete and in-depth understanding of that film. Since Bardo is an Alejandro G. Iñárritu film, any claim to have grasped its every nuance and implication is downright daring, so we’ll leave that as an exercise for our more cinematically intrepid readership.
Selecting from such a long and distinguished credit history as Khondji’s will always risk omitting someone’s favorite. Let’s at least mention Delicatessen, Amour, and the mighty Seven. The latter is probably up there with Saving Private Ryan and Blade Runner as a common inspiration for overambitious microbudget directors. Bardo looks nothing like those films, but it is filled with moments of spectacular surrealism, making it another tricky creative reference for small productions.
The film has amazing sets—trains full of water, houses full of sand. Shooting good-looking things is, after all, a reliable way to make things look good. Bardo’s particular genius, though, is in surprising the viewer by presenting what looks like a straightforward scene, then subverting that straightforwardness by filling the frame with things that are entirely outside the context of normal human experience.
That said, there’s one consistent and noticeable choice that’s available to more or less everyone, and that’s shooting wide. Wide lenses didn’t always look the way they do in Bardo, at least not unless you’re shooting 65mm film. Bardo was shot on Alexa 65s, which may not be the least expensive camera system around but is at least a little less expensive than thousand-dollar rolls of 65mm negative.
What’s unusual is that most people will associate the larger sensor with shallow depth of field. That’s not wrong; as we saw with All Quiet on the Western Front, the ability of big chips to render shallow focus even at wide angles is a wonderful thing.
That’s not the choice made for Bardo, though. The film shoots places and people like a stills specialist might shoot architecture. Things which are nearby loom out at us; low angles look really low, and make high things look really high. The scale difference between foreground and background is maximized. Anything that’s close is really close, and anything that isn’t might as well be in the next time zone.
That much might already be clear to anyone who’s ever shot stills of buildings for people who sell buildings. In motion, there’s more. A wide-angle lens can, if we want it to, make moving objects blast past the viewer with almost fighter-jet levels of energy.
Differences in perceived speed pic.twitter.com/V42VxQDRZ3
— Akiyoshi Kitaoka (@AkiyoshiKitaoka) February 4, 2020
Often, all of these things are used to emphasize characters over the background, as well as just leveraging the wide field of view to give us a good look around a scene without having to cut or move.
In some ways, that’s also the downside, at least for smaller shows. Less well-funded productions than Bardo often live and die on their ability to exclude various parts of a scene. There’s often a need to avoid an inadequate bit of production design or poorly-placed fire exit sign in a period drama.
“In some ways, that’s also the downside, at least for smaller shows.”
Wide angles create much the same issue as things like Steadicam and gimbals in the wrong place. It’s great, but it’s very easy to show too much, too fast. Small productions can struggle to keep up.
Short films on tiny budgets were actually among the first to start shooting full frame, back in 2009 when Canon released the 5D Mk. II. Today, anyone with a few thousand to spend can afford a full-frame camera that’s more than capable of results that a mainstream cinema audience wouldn’t question. Many of them have shallow enough mounts to accept almost any stills photo lens. It all works—but it has to be done with care.
Obviously, it’s unfair to boil down such an accomplished film just to its framing. Khondji uses a huge raft of advanced techniques, with LED video walls to simulate a moving train, day-for-night setups, and the sort of day interiors where the audience can feel what the weather is like outside even though it’s actually a couple of 18Ks and a SkyPanel. Crucially, making the weird seem normal allows the surrealism to creep up on the audience.
Bardo’s approach to lensing is probably unusual enough to be worth singling out, especially as it’s something that people at the beginning of their careers often overlook. It’s too easy to simply bounce in on a zoom and put an actor in the right place in frame. What we sometimes miss is where that puts everything else. If Bardo encourages more people to think about the other options, whether that involves a wide lens or not, then that’s a good thing.
Cinematography by Mandy Walker, ASC, ACS
Mandy Walker and Baz Luhrmann have collaborated on several previous occasions. The feature Australia is one example, but perhaps more visually relevant to Elvis are two Chanel commercials from 2004 and 2014. So famed are they for their gorgeousness that both are still attracting admiring YouTube comments, one almost twenty years after it was made.
Those, and much of Luhrmann’s work, might lead us to expect colorful, intense, downright theatrical pictures. And Elvis does not disappoint, though it’s absolutely not an extended music video. The film opens in the dusty environs of a practically sepia-tinted 1940s South. Scenes bathe in a glow of tungsten and occasionally harsh sunlight, but it’s a warm world overall.
Early musical performances are resplendent in the kind of clean-cut 1950s aesthetic that Presley’s early persona seemed dedicated to gleefully disrupting. Things drift toward the lambent cyanotic gloom of 1970s fluorescents as things begin to go wrong, though it’s a realistically mixed world throughout.
Whether or not anyone’s enough of a Presley historian to notice, some of the key performance scenes are painstaking reconstructions of historic broadcasts. There’s lots of fun to be had here with period lighting techniques. Does anyone use footlights anymore?
In 2023, a pronounced uplight is more likely to be interpreted as a slightly mid-century Hammer-horror attempt to make something look spooky. In Elvis, though, a long row of it becomes a time machine to a mid-century look that really puts the show into show business. Many of those scenes reportedly involved sourcing historic lighting gear to suit several different time periods.
Unlike some of the films we’re discussing, though, Elvis is not easily associated with a fixed palette. Gel manufacturers have occasionally described some of their selections as “party colors,” and it might be instinctive to look there in search of this kind of live-performance aesthetic. Historically, people did exactly that, especially at a time when color television was new and special and the money people liked to see color used.
Elvis isn’t really even a party-colored film, though. There’s plenty of pink and gold throughout, but the performance scenes are often less than primary-colored. LEDs make it easy to reach for the harshly-saturated shades which are likely to define post-2000 club and performance lighting.
Mid-century lighting directors might have struggled to create those colors, particularly super-deep blues, even if the cameras could have dealt with them. So, at a mechanical level, Elvis advises us not to turn the saturation control on our color-mixing LEDs all the way up to eleven.
The latter parts of the film look to have been photographed with true anamorphic lenses (the Panavision T series, according to some sources). The entire production is in a cinemascope-esque aspect ratio, but a close look at focus artifacts quickly reveals that earlier scenes were shot with spherical glass (the Sphero 65, if online sources are to be believed). The result is that the world blurs around the edges as the character does much the same.
But our purpose here is to find ways low-budget filmmakers can take inspiration from the high end, which might seem tricky. The film captures upscale performances in spectacular locations, with hundreds of extras in period costume, with multiple cameras. That’s before we’ve even considered the work involved in shooting musical performances with a fleet of Alexa 65s with multiple sets of 65mm-compatible Panavision lenses.
This is not a simple nor an inexpensive way to make a movie.
“What Elvis demonstrates, to absolutely spectacular effect, is how things can be consistent without being identical.”
What it best tells us about, perhaps, is progression. It’s easy to be sucked into the idea of defining a look for a production then dedicating oneself to maintaining that look, and only that look. Consistency is, of course, key in dramatic camerawork. What Elvis demonstrates, to absolutely spectacular effect, is how things can be consistent without being identical. It shows us how cinematography can shift to support the changing mood of a story without making an end that looks like a different movie than the beginning.
All of this is difficult. It requires a massive amount of control and planning. Many of the things that small independent filmmakers can take from the high end are comparatively simple techniques that can be straightforwardly applied to almost any production. Use backlight; use color consistently, use atmospheric effects, use interesting framing. What we get from Elvis is “use a carefully-considered progression of photographic technique to reflect and enhance the changing narrative, without risking jarring inconsistency.” There isn’t really any easier way to break that down.
There are a million ways to do that, and it’s potentially zero-cost, depending how it’s executed. It’s just the sort of zero-cost concept that requires practicing for decades.
Empire of Light
Cinematography by Roger Deakins, ASC, BSC
It surprises many people to know that Deakins’ career went without Academy recognition until the award for Blade Runner 2049 in 2018. He had already won three BAFTAs by that point (for The Man Who Wasn’t There, No Country for Old Men and True Grit) and collected two more for the Blade Runner sequel and for 1917, which would also be recognised with a second statue from the Academy.
Mendes and Deakins have collaborated before, on 1917 as well as Skyfall and Jarhead. Those films are all wildly different in look and tone, but just as spectacular. So, while Empire of Light might instinctively seem like a modest little story, it might be a mistake to limit our expectations.
Set in the early 1980s, it follows the manager of a seafront cinema in southeastern England, her friendship with a new employee, and the way that relationship interacts with her mental health problems and outlook on life.
It is a powerful character piece, with one of those characters at least arguably being the titular Empire Cinema. It’s even the first thing to feature in the trailer, when the art deco interior springs to life in spectacular bursts of golden light.
It’s a theme which continues through the production. It’s been suggested that Deakins deliberately specializes in exactly that kind of soft, lambent illumination, though that seems a bit reductive in the context of things like Jarhead. Certainly there are superb examples of his work (particularly Blade Runner 2049, and the interiors for True Grit) where we might be forgiven for forming that impression, and it’s certainly the signature look for a lot of Empire of Light.
So let’s talk about soft light, if only because some people have the habit of hanging a pop-up softbox on an LED then putting it so far away that it barely matters.
The basics are easy. Larger, closer lights are more soft, and smaller, more distant lights are less soft. For an example at the upper end of the range, look at the closeups of Nicole in the Baz Luhrmann Chanel commercial we mentioned earlier. In that case, “closer” probably means barely out of frame, and “larger” probably means ten feet square. The advantage for a commercial, shot as a montage, is that it probably only needed to work for small, brief closeups.
Long dramatic dialogue scenes where the actors move around a lot are trickier. Trying to give a rule of thumb about sizing soft light sources invites uproar in the comments. Let’s say anything claiming to be a soft light will need to be about as wide and tall as the distance between itself and the subject.
That distance may be quite large for very mobile scenes, so the soft light might end up being really, really large; really big setups might use diffusion tens of feet on a side, or huge floating helium balloons. Car commercials often use enormous overhead white panels which are practically part of the studio.
We can safely imagine the lighting people emptying the truck of every piece of gear on hand in order to create and, moreso, control the sort of soft sources that seem to illuminate much of Empire of Light. That’s not the case for absolutely every single scene, of course. Things get a little harsher and less comfortable as the lead character’s mental state deteriorates. A pivotal scene in the film, though, is certainly lit with bounced, diffused light—the light of a cinema screen, illuminated only by the projector.
Scowling greybeards have sometimes complained that soft light is cheating. Back in the days when film stock had a tenth the sensitivity, illumination had to be applied at storm force. Look at black-and-white movies of the 1930s and 1940s, which often hit people right in the face with powerful, un-diffused light sources. That requires people and cameras to hit marks accurately or it looks terrible.
Softer light potentially creates a little flexibility. As a practical matter, though, especially on a smaller show, it’s more than offset by the challenges of containing a large light source which can be very hard to control. That’s something this film does wonderfully.
For the rest of us, it means becoming familiar not only with eight-by-eight diffusion, but also black solids of about the same size. There are a million things to learn from Deakins, but anyone who’s interested in looking at advanced techniques with soft sources need look no further than Empire of Light.
Cinematography by Florian Hoffmeister, BSC
Relying on production design is easy when we’re discussing an epic period piece or spectacular sci-fi. On the other hand, placing a story in a modern city, in apartments, streets, and performance venues, might seem like less of an opportunity to shine. With that in mind, let’s recognise that the production design of Tár is utterly gorgeous.
The film is set in clean, angular environments. It’s also rendered in the sort of muted, thoroughly modern color scheme a contemporary, high-end interior designer might select as a mood board swatch.
If we hadn’t already talked about locations and color control, that might be our subject here. Since we have, we’re going to concentrate on something which is at least as big a part of Hoffmeister’s work on the film. We could easily talk about blocking, framing, and the use of a limited number of longer setups, but that doesn’t really go far enough. Much of Tár is practically an exercise in geometry, especially those parts of it which take place in architecturally interesting environments.
One worthwhile exercise is separating the look of Bardo, which glories in its wide lenses, from the look of Tár, which integrates the camera very differently with its surroundings.
The film was shot using larger-sensored cameras, but not every scene used the full area, even taking into the 2.39:1 frame. Deeper depth of field was specifically chosen for certain scenes. ARRI’s system of replaceable rear elements for the Signature Primes was used to bring some variation, though even that seems to have been avoided in the darker scenes, where any reduction in contrast and sharpness risks distraction.
Despite the high gloss of Tár, this kind of photography is potentially quite accessible to smaller shows. Words like clean and precise also describe the lenses we often get on modern stills cameras—almost to a fault, in fact, which is where things like filtration come in.
Or, go to eBay and wince at the prices of things like Asahi Takumars, if they can be adapted to your mirrorless camera of choice. They don’t look anything like Signatures, but closed down a little, they can be clean without being harsh.
The thing Tár specifically teaches us is about integrating lens choices with blocking and staging. Often, student films are way too—what’s the word—cutty, because it’s often easier to shoot many setups and chop between them. Conversely, it’s increasingly common for low-budget filmmakers to leverage the accessibility of gimbals to create long, complicated, conspicuously clever single-setup scenes.
Hoffmeister and Field often do neither of those things, relying on a single, simple, setup with limited mobility, and staging things to suit.
Tár is an object lesson in this, and it’s a worthwhile exercise to watch the thing with a notepad in hand and figure out exactly how many lenses were used just by looking. Of course, there are a lot of extra complexities besides that.
The film is about a composer and conductor, which means that it faces some of the same practical challenges as Elvis as some scenes become effectively a music video. The costs of hiring a full symphony orchestra adds time pressure. The need to have story-relevant things happen at the right moment in the music is a further complication. All the while Cate Blanchett is acting a hugely specialist task while also appropriately playing a character.
This is Hoffmeister’s first Oscar nomination, and his first collaboration with director Todd Field. In interviews, the two talk enthusiastically about their decision making process with regard to blocking, framing and selecting lenses. When people are practically finishing each other’s sentences like that, it’s hardly useful to divide up the creative responsibilities, but there are inclinations toward this kind of careful framing in Hoffmeister’s other work. It’s visible in things like Official Secrets and the Apple TV series Pachinko, though less so than Tár.
Put all these things together, and the result is something that’s simultaneously profound and straightforward. Different people have different ideas about what the core purpose of cinematography is or should be. Probably we can agree that, like sound design, it should mostly do something powerful and subtle while simultaneously sliding past the conscious attention of the viewer.
It should be there if we look for it—and it is, in spades—but not otherwise intrude. If there were an Oscar for simultaneously being both unshowy and spectacular, Tár would be a shoo-in.