What is a lighting check: the term explained
A lighting check is a capture-quality control that tests whether there is enough usable light on the subject before or during recording, so the person reviewing the result later can tell what is actually in the frame. Too dark, blown out or lit from the wrong side, and the step is flagged for a redo.
It is one of the least glamorous parts of collecting visual information remotely, and one of the most decisive. A reviewer cannot assess what they cannot see.
What does a lighting check actually measure?
Three separate things, which is why "it was too dark" is usually the wrong diagnosis.
- Overall brightness. Is enough light reaching the sensor at all? A phone compensates by raising ISO, which buys brightness and pays for it in noise.
- Dynamic range. A dark panel next to a bright window forces the camera to pick one. Whichever it drops becomes unreadable.
- Direction and reflection. A ceiling strip light bouncing off wet paint hides precisely the scratch someone was asked to record.
That third one, specular reflection, is the most common reason a technically well exposed photograph is still useless on an inspection.
How much light is enough?
There is no universal threshold, but the workplace lighting standards give a reference point worth knowing. EN 12464-1, the European standard for lighting of indoor work places, sets 500 lux as the maintained illuminance for detail focused visual tasks such as reading, writing and close inspection work. A properly lit workshop bay clears that comfortably. An underground car park at night does not come close, and neither does a driveway at seven on a November evening.
Phones make this harder to judge by eye. The screen preview is brightened and stabilised before anyone sees it, so a frame that looked fine while recording can arrive as a grey smear.
How is a lighting check run?
Two approaches, usually combined.
- Automatic. The software samples the frame, estimates exposure and contrast, then warns or blocks before the step is accepted. Fast and consistent, and blind to context: it cannot tell that the bright patch worrying it is the sky.
- Instructional. The step itself tells the person what to do. "Stand with the daylight behind you" prevents more bad frames than any warning shown afterwards.
Where the check fires matters as much as how it works. A prompt during the session costs the participant ten seconds. The same problem found the next morning costs a retake request, a reply, and a day of cycle time.
Lighting check example: a scratch in an underground car park
A driver is asked to record damage to the nearside door. He parks under the ramp, where the only fitting is six metres behind him, and records his own shadow across the panel.
An automatic check flags the frame as underexposed, and the step repeats with one added instruction: use the phone torch, held at an angle to the panel rather than straight at it. The second attempt shows the scratch and its depth. Same driver, same phone, same car, four minutes later.
What a lighting check will not fix
- It is not blur detection. Blur is about focus and motion. A perfectly lit frame can still be shaken, which is why most quality layers run both tests.
- It does not judge coverage. A bright, sharp photograph of the wrong wheel passes every exposure test there is.
- It cannot rescue a dark original. Brightening afterwards raises the noise along with the signal. The detail was never recorded in the first place.
- It says nothing about authenticity. Lighting quality is unrelated to whether the scene is genuine. That question belongs to upload versus capture and the evidence layer around it.
Designing for light instead of checking for it
Teams that get consistently usable submissions do most of the work upstream, in the capture flow rather than in the warning dialog.
- Say where to stand relative to the light, not only what to point at.
- Prompt the torch explicitly for engine bays, wheel arches, lofts, cupboards and anything under a sink.
- Ask for wet surfaces to be dried, or accept that reflections will be part of the record.
- Where the assessment depends on colour, say so, and expect worse results under sodium and older fluorescent lighting.
- Keep a plain fallback. If the light genuinely cannot be fixed, let the participant say so rather than abandon the request.
None of that is exciting work. It is also the difference between a guided photo capture flow that returns decisions and one that returns another round of questions.