Flicker is invisible to the eye: it shows up in the videos your marketing shoots

Light and visual comfort

A lobby that looks perfect in person can be unusable on camera. It is a fault no commissioning catches, because commissioning is done with the eyes and the eyes do not pick it up.

A light that switches off a hundred times a second

An LED does not emit a continuous light: it emits pulses, because the power supply turns alternating current into a current that pulses. When that pulsing is deep and slow it is called flicker, and its depth depends almost entirely on the quality of the power supply, not on the LED.

The human eye averages: above a certain frequency it no longer sees the individual pulses and reads a steady light. That is why a room can look perfect to whoever is standing in it and not be perfect at all. The visual system, however, keeps receiving the oscillation even when it does not recognise it, and that is where the two effects we will discuss come from: the one on people and the one on cameras. And there is a practical consequence that surprises people: two fittings with exactly the same source can behave in opposite ways in front of a camera, because what counts is not the diode but the electronics feeding it.

The camera does not average: it samples

A digital sensor does not look continuously: it opens and closes a shutter many times a second. When the frequency of the light and the frequency of the shutter do not agree, the result is visible and unmistakable — horizontal bands rolling across the image, or a shot that changes brightness from one frame to the next.

In a hotel this is not a technical problem: it is a commercial one. Marketing shoots for the website, the couple shoots at the reception, guests shoot on their phones, and so does anyone producing content in the room you have just finished lighting. A room that pulses on video does not get shown, and it does not get shown at exactly the moment you would need it most. And it is a fault that spreads: once a pulsing clip is online it stays online, and whoever watches it does not think «they got the lighting wrong», they think the place simply looks like that.

Slow motion makes it worse

Slow-motion footage, the kind used in almost every hotel promotional video, increases the number of samples per second. The more samples there are, the more finely the camera reads the pulsing, and a fault that was barely perceptible at normal speed becomes obvious.

That is why some rooms turn out inexplicably hard to film: the videographer tries, it does not work, changes settings, and in the end brings their own lights and switches yours off. That moment is the rejection of the supply, and nobody writes it in a report — but it comes back when the supplier for the next property is chosen.

The five places where it really costs

They are not the brightest places: they are the ones that end up in frame. In a hospitality property the value of a room is also measured in how often it gets photographed, and some rooms exist almost only for that:

• the event and wedding room, where every shot is content that circulates for years
• the lobby and reception, which are the first and last shot of almost every video
• the restaurant, where the plate is photographed by the guest and the colour has to hold
• the spa and the pool, filmed in low light: that is where deep dimming makes the fault worse
• the façade at night, which ends up in every arrival photo and every listing page

The effect on people, which is documented

Above the threshold of conscious perception flicker does not disappear: it keeps reaching the visual system. Studies associate it with headaches, fatigue and — in a sensitive minority — with more serious disturbances. In an office the discussion is about productivity; in a hotel it is about a guest who sleeps badly and does not know why.

It is not something to put on the front page of a brochure, and it should not be dramatised. But it is the reason why standards on visual comfort in workplaces have begun to name it explicitly, and why a well-written specification today lists it next to glare instead of ignoring it. It is worth asking the supplier precisely because it is awkward: the answer separates those who have measured from those who repeat what is written on their own supplier's catalogue.

Two numbers, and what they really say

The first is the percentage of oscillation: how far the light drops between one pulse and the next, from zero to a hundred. Zero is a steady light, a hundred is a light that goes fully out on every cycle. The second is an index that weighs the depth together with the frequency, because a fast oscillation disturbs far less than a slow one of the same depth.

A serious supplier publishes both, or at least the first, and publishes them for the driver actually fitted, not for the family. A supplier who does not have them has probably never measured, and in that case the honest answer is «I do not know»: that is useful information too, because it says the risk is yours to carry.

What to write in the specification

There are four lines, and they belong in the rooms that get filmed, not everywhere: demanding television-studio performance in a warehouse is a way to push the price up without buying anything.

• the maximum oscillation percentage allowed, measured at a hundred per cent output
• the same percentage measured at the lowest level the evening scenes foresee
• the index that weighs depth and frequency, with the required value written out in full
• who runs the filming test on site, with which camera and at what slow-motion rate

The fault gets worse as you dim

This is the point almost no specification covers. Many cheap drivers behave acceptably at full output and badly when dimmed: the modulation deepens exactly as the light goes down, which is to say in precisely the evening scenes a hotel uses most.

Asking for the figure only at a hundred per cent is therefore asking for the figure in the easiest case. The measurement that matters is the one at the level you will actually use at eleven at night in a lobby: if the supplier provides it, that is a supplier who has measured; if they provide it and the number barely worsens, that is a supplier who chose a good power supply. In practice there is one question to ask, and it is very concrete: what is the oscillation at ten per cent output? If the number does not exist, the evening scene is a gamble.

It is not repaired afterwards, it is replaced

Flicker is born in the power supply, so in theory it is cured by changing the power supply. In practice, on a bespoke decorative fitting, the supply sits inside the body and the body was designed around it: changing it means reopening the piece, and sometimes remaking it.

On technical fittings with a remote driver, replacement is possible and costs little in materials, but it has to be done on every fitting in the room, not on one: two different drivers in the same room give two different behaviours, and the camera's eye separates them instantly.

How to test it, in half an hour

You do not need a laboratory instrument to notice. You need a phone, an improvised tripod and the slow-motion function any recent phone has. Switch the room to the scene you plan to use, film a light-coloured surface for ten seconds and watch the clip on a big screen.

If bands appear and roll, the fault is there. The test gives no number and does not replace a measurement, but it answers the question that matters — is it visible or not — and answers it before the order goes out, when changing it costs an argument instead of a building site. It is also the quickest way to get the designer and the supplier to agree, because it moves the discussion off opinions and onto a video watched together.

Why it always ends up outside the specification

Because it does not show at commissioning. A lighting handover checks levels, uniformity, glare and consumption: all things measured with the eye or with a lux meter. Flicker enters none of those measurements, so it passes, and it passes with everybody's signature on it.

It shows up months later, when somebody shoots a video, and by then there is no longer a moment in which to raise it: the supply has been accepted, the balance has been paid, and the fault has the air of a whim of whoever is filming. It is the classic defect prevented in one line and never repaired in a negotiation. Prevention costs a line of specification and half a day of testing; repair costs replacing the fittings of an entire room, and meanwhile the room cannot be used.

How we handle it

On bespoke pieces we choose the power supply before closing the drawing, because a low-flicker driver takes more room and heats differently: choose it afterwards and it does not fit, and what fits is the small one. It is one of those decisions that look technical and are decisions of design.

And when a supply concerns rooms that will be filmed, we say so in the quotation instead of waiting to be asked: almost nobody asks, because almost nobody knows it can be asked, and a supplier who stays quiet about this is selling a deferred problem.

• A bespoke piece
• The properties we have lit
• Free lighting calculation
• Expert advice