What it buys you
Audiences read light for plausibility without knowing they are doing it. A frame lit from a direction with no visible cause feels staged, even when it looks nice. Put a lamp in the corner and light the room from that corner, and the whole shot becomes credible, because the eye finds the answer to a question it asked automatically.
That is the first function of a practical light: it explains the rest of the setup. The second is more literal. Sensors got fast enough that a domestic bulb is now a usable key, so the fixtures in frame can genuinely light the scene rather than merely justify it. A great deal of current film and television is lit almost entirely this way, which is why so many night interiors have visible lamps in the composition.
Rigging them so they work
The fixtures a set decorator provides are rarely usable as they arrive. The standard interventions:
- Change the bulb. Higher output, and a colour temperature you choose rather than inherit. Mixed colour across practicals is the most common reason a room looks dirty rather than warm.
- Dim everything. Every practical on its own dimmer, so the balance between fixtures becomes a decision you can make from the monitor instead of a fixed accident.
- Hide light inside. LED tape or a small panel inside a shade, behind a translucent screen, or under a table edge. The visible fixture reads as the source while the hidden unit supplies the level.
- Add a cheat just out of frame. A small warm source next to a candle or a phone screen, matched in colour, so the face gets exposure while the visible flame or screen keeps the credit.
- Watch for flicker. Test any LED or fluorescent unit at the frame rate and shutter you will actually shoot, before the day.
Exposure is the judgement call. A practical that sits at the same level as the face reads as switched off. One that sits four stops over turns into a featureless white shape and pulls the eye out of the frame. One to two stops over is the range where a lamp looks lit and still holds detail.
The look it produces
Because these sources are physically inside the set, they behave like real ones. Falloff is steep, so a subject leaning toward a lamp gains a stop and leaning away loses one, and that change happens on camera as the actor moves. Colour is warm and often mixed. Pools of light appear between dark gaps rather than an even wash.
All of that is why practical-driven lighting reads as observed rather than arranged, and it is also why it is harder to control. Continuity across coverage takes discipline: the lamp is in the wide, so it has to be in the same state and the same relationship to faces in every angle.
Prompting a practical light in AI generation
This is the most reliable lighting instruction available in a prompt, and it is underused. Models are strong at rendering a named fixture, and once the fixture is in the frame they tend to light the subject consistently with it, because that pairing appears constantly in their training data. Naming a lamp fixes direction more dependably than naming an angle does.
- ✅
a single shaded desk lamp at frame left as the only source, warm pool of light on the wall, face lit from that direction, ceiling dark - ✅
neon sign outside the window as the only source, magenta light across one side of the face - ⚠️
practical lighton its own means little to a model, which needs to know which fixture - ⚠️
lamp in the shot, dramatic side lighting from the rightsets up a contradiction if the lamp is on the left
Two habits are worth keeping. Say what the practical does to a surface, a pool on the desk, a gradient up the wall, a stripe on the floor, because that is what forces the model to make the light and the fixture agree. And in video, say the source is steady unless you want flicker: an unconstrained model will often pulse a lamp or animate a television's content in ways that read as an artifact rather than as a choice.