How to Film LED Walls and Screens Without Flicker, Banding or Moiré

Walk onto any conference stage, trade-show floor or product launch today and you are standing inside a light source. Keynote backdrops are LED walls. Booth architecture is built from LED panels. Speaker support runs on confidence monitors, and half the B-roll a client wants involves someone looking at a screen. The modern corporate environment is made of displays, and every one of them is capable of ruining a shot in a way the eye never sees on the day.

The failure modes are familiar to anyone who has reviewed conference footage: dark bands crawling across the keynote screen behind a CEO, a booth wall that pulses between frames, a rainbow shimmer across a presentation slide that was perfectly clean to everyone in the room. None of it is visible on set without knowing where to look, and almost none of it can be repaired in post. Filming screens well is a discipline of prevention, and it starts with understanding why displays and cameras disagree in the first place.

Why Screens Misbehave on Camera

An LED wall does not emit light continuously. Each LED is switched on and off thousands of times per second, and brightness is controlled by pulse-width modulation — varying how long the LED stays on within each cycle rather than how much current it draws. On top of that, most walls are multiplexed: the panel drives its rows of LEDs in sequence rather than all at once, scanning through the display faster than the eye can perceive. To a person in the room, the result is a stable, bright image. To a camera sampling the wall in discrete slices of time, it is a strobe.

How badly the strobe shows depends on the wall's refresh rate. Broadcast-grade panels refresh at 3840Hz or higher, which most cameras integrate cleanly at normal shutter speeds. Event and rental walls are frequently 1920Hz, and older or cheaper panels can be lower still. The difficulty at a conference is that the wall is never your choice. The AV contractor installed it days before you arrived, the production office does not know its refresh rate and the panels may even be mixed — two hire stocks bolted into one backdrop, refreshing at different rates. You film what is there.

The second variable is the camera, and specifically how its shutter samples time.

What a Rolling Shutter Does to a Screen

Most cameras on corporate shoots — mirrorless hybrids, camcorders, nearly all drones — use rolling shutter sensors. The sensor does not capture the whole frame at once. It reads out line by line, top to bottom, so the top of the frame is exposed at a slightly different moment from the bottom.

Against a static subject this is mostly harmless. Against a pulsing light source it is the direct cause of banding. Because each row of the sensor samples a different point in the wall's refresh cycle, some rows catch the LEDs mid-pulse and others catch them between pulses. The result is horizontal bands of varying brightness across the screen — and because the camera and the wall are not synchronised, the bands crawl. A static defect might pass unnoticed. A moving one draws the eye instantly, and it sits directly behind the speaker in every keynote wide.

There are partial workarounds on a rolling shutter camera: adjusting shutter speed until the bands slow or thin, hunting for the least-bad compromise between exposure and artefact. But it is a negotiation, and on a live stage with one take of a keynote, a negotiation you can lose.

What a Global Shutter Changes

My A-camera, the RED Komodo, uses a global shutter sensor: every photosite on the Super 35 sensor begins and ends its exposure at the same instant. There is no line-by-line readout during capture, so there is no mechanism for different parts of the frame to sample different moments of the wall's refresh cycle. Intra-frame banding — the crawling stripes that define rolling shutter footage of screens — is structurally impossible.

That is the honest scope of the advantage, and it is worth being precise about what remains. A global shutter samples the whole frame at one moment, but successive frames can still land at different points in the wall's pulse cycle. If the exposure time is not a clean multiple of the refresh period, the wall's brightness varies slightly from frame to frame — a uniform, whole-frame flicker rather than bands. The difference in practice is that whole-frame flicker is predictable and removable at the source: adjust the shutter until each exposure integrates the same number of refresh cycles, and the wall steadies. A crawling band on a rolling shutter is a moving target; flicker on a global shutter is a tuning problem with a stable solution.

The Komodo helps here in a second way. Its shutter can be set in precise time increments rather than a short menu of standard speeds, which means I can walk the exposure in small steps while watching the wall until it locks. At normal production shutter speeds — 1/50 at 25fps, say — a 20-millisecond exposure spans dozens of refresh cycles on even a 1920Hz wall, which is why a properly set shutter renders most walls cleanly. The problems concentrate at short exposures and high frame rates, where each frame integrates only a handful of cycles and the wall's multiplexed scanning can surface as visible dark structure. If a client wants slow motion in front of an LED backdrop, that is a conversation to have before the shoot, not a discovery to make in the edit.

There is one more screen-adjacent benefit that matters on stages. Moving lights, LED fixtures and camera flashes from press photographers all pulse or fire mid-frame. On a rolling shutter, a photographer's flash exposes only the rows being read at that instant, leaving a bright band across part of the frame. On a global shutter the flash raises the whole frame evenly — one bright frame, easily managed, instead of a torn one.

What the Shutter Does Not Fix

A global shutter solves the temporal problems. It does nothing for the spatial and colour problems, and pretending otherwise is how screens get ruined in a different way.

Moiré is the first. An LED wall is a grid of discrete emitters, and a sensor is a grid of discrete photosites. When the two grids interfere — when the wall's pixel pitch, the shooting distance and the focal length combine so that the panel's structure lands near the sensor's resolution limit — the result is shimmering interference patterns and false colour across the screen. This is optics, not timing, and no shutter design touches it. The working controls are distance, angle and focus: step back or go longer until the panel's pixel structure dissolves below what the lens resolves, avoid straight-on framings at the critical distance and, where the shot allows, let the wall sit slightly behind the plane of focus. A speaker sharp against a marginally soft backdrop reads as depth. A speaker sharp against a moiré pattern reads as a mistake. Moiré also cannot be fixed in post, so I check for it at the actual shooting position, on a proper monitor, before the session starts — it can appear at one distance and vanish two metres back.

Exposure balance is the second. An LED wall is self-emissive and, at typical event brightness, far brighter than a lit human face. Expose for the speaker and the wall clips; expose for the wall and the speaker sinks. The camera's dynamic range buys room to grade, but the real fix happens before the lens: I ask the AV desk to bring the wall's brightness down for camera, often substantially, because walls are routinely run at levels set for the back row of a hall rather than for a sensor three metres from the stage. Most operators will do it in seconds if asked — and almost nobody asks. Then I light the speaker to hold their level against the wall, which matters doubly on the Komodo because I treat ISO 800 as its working ceiling: the face has to be genuinely lit, not rescued in the grade. One warning from experience: LED panels can behave worse at low brightness, because dimming happens through the same pulse-width modulation that causes flicker in the first place, and some panels lose grayscale stability when driven low. After any brightness change I re-check the wall on camera before rolling.

Colour is the third. LED walls produce white from narrow bands of red, green and blue, which is nothing like the continuous spectrum of daylight or tungsten. A wall that looks neutral to the eye can record with a magenta or green cast, and its rendering shifts with viewing angle. Where the screen content matters — brand colours, product imagery — I check it on a calibrated monitor at the shooting position and correct at the source or in white balance, not on faith. And I ask what frame rate the wall's processor is outputting: content playing at 60Hz while I shoot 25fps introduces its own cadence problems that no shutter setting removes.

How This Plays Out on a Live Stage

Conference coverage is where all of this converges, because a keynote offers no second take. I have filmed multi-day conference programmes for clients including Huawei at Web Summit, the World Aviation Festival and EUPVSEC in Lisbon — environments where the entire stage picture is an LED surface and the schedule does not pause for a camera department to experiment.

The routine that works is front-loaded. During rehearsal or the pre-session lull I put the camera at its actual positions, frame the wall, and check three things in order: shutter, until the wall holds steady frame to frame; moiré, at every focal length I expect to use; and balance, negotiating wall brightness with the AV desk while there is still time to relight the lectern. It takes twenty minutes and it is the difference between a keynote backdrop that looks like a broadcast and one that flickers behind every soundbite the client wants to cut.

The same discipline transfers to smaller screens. Laptop displays, tablet close-ups and control-room monitors in industrial B-roll all pulse in their own ways, and the approach is identical: sample the screen through the camera before committing to the shot, adjust shutter until it settles, and manage brightness at the device rather than in the grade.

The Case for Getting It Right at Capture

Everything in this article shares one property: it is cheap to solve on set and expensive or impossible to solve afterwards. Banding baked into a rolling shutter keynote cannot be removed. Moiré cannot be filtered out without destroying the screen content it sits on. A clipped LED wall holds no detail to recover. The screens in a corporate video usually carry the client's own message — their slides, their brand, their product — which makes them precisely the element that cannot be quietly degraded.

A global shutter camera removes the worst and least forgiving of these failure modes outright, and disciplined technique handles the rest. That combination is why I shoot stages, screens and display-heavy environments the way I do — and if your next production involves a keynote, a booth or anything else built from light, I would be glad to talk about it.