Some LED lights switch their output rapidly to control brightness, a method known as pulse-width modulation; others have small fluctuations from their power supply. Your eyes may blend the pulses into apparently steady light, while a camera samples the scene during short exposures. If the sampling and light timing do not line up, video can show flicker, pulsing or horizontal bands.
A steady-looking lamp may be switching
LEDs can be dimmed by rapidly switching them on and off, changing the fraction of each cycle when they are lit. The pulses may be too fast for a person to notice directly, but the light output still varies over time.
The camera samples a different moment
A camera does not average light exactly like human vision. Its shutter sets an exposure window, and many digital cameras read image rows at slightly different times. When the exposure or rolling-shutter scan interacts with the LED’s pulse rate, different frames or rows capture different brightness. The result can be flashing or moving dark bands even though the lamp looks steady in the room.
This is a timing interaction. It does not necessarily mean the lamp is visibly strobing to people, though some LED flicker can be noticeable without a camera.
Ways to reduce the artifact
Try a different shutter speed or frame rate, use the camera’s anti-flicker setting, or raise the light’s brightness if its dimmer uses a slower pulse pattern. For filming, test the exact lamp and camera combination before recording. Professional fixtures may offer high-frequency or flicker-free modes.
A useful check is to compare the scene directly with the recorded image. If the bands appear only in the footage, the problem is likely the interaction between light modulation and camera timing.
Sources and evidence
- MIT Computational Photography — Image sensors and rolling shutter ↗
How rolling-shutter rows can record different phases of flickering light.
- IEEE Spectrum — LED lighting and camera flicker overview ↗
How LED modulation can appear as a camera imaging artifact and the IEEE P2020 measurement context.
Sources consulted: 11 October 2026. Written with AI assistance and checked against the linked references; not independently reviewed by a subject specialist. Sources do not endorse this article. See our editorial policy.
Revision: Original explanatory article prepared from linked scientific and institutional sources; scope and uncertainty are stated in the text.

