How LED video displays work

2.1 Additive RGB colour mixing

The precise spectral output of the LEDs matters. Two nominally red devices can have different dominant wavelengths and spectral widths. Component selection, binning, drive conditions and colour management therefore influence the achievable colour gamut and the consistency between pixels, modules and cabinets.

2.2 Pulse-width modulation and greyscale

LED output is commonly controlled through rapid switching and current regulation. Pulse-width modulation varies the proportion of time for which an LED is active within a cycle. Driver electronics also govern row scanning, current and timing. The viewer normally perceives the average light output rather than the individual pulses.

This interaction is one reason that refresh rate, scan design, driver performance, low-level greyscale and camera shutter settings cannot be considered independently. A display may look stable to the eye but show bands or artefacts on camera. Conversely, a high headline refresh rate does not guarantee excellent low-brightness performance.

The US Department of Energy's explanation of flicker is written for lighting but helps explain why modulation frequency, depth and viewing conditions all matter. Video displays add further timing relationships between content frame rate, processing, scanning and camera capture.

2.3 Modules, cabinets and mapping

A module contains a defined number of pixels. Cabinets combine modules with receiving electronics, power distribution and mechanical features. The display is then mapped so that the video processor knows which portion of the image belongs to each cabinet and module.

This mapping allows a non-standard canvas to be treated as one coordinated image area. It also means that the source content should match the actual pixel dimensions of the display wherever practical. Scaling can be excellent, but native-resolution content gives designers the most direct control over line weight, text size and detail.

2.4 The high-level signal path

A typical system can be understood as a chain from content source to light output:

1. A computer, camera, media server or signage player supplies video.

2. Image processing scales, switches, crops or maps that content.

3. A display controller converts the image into display-specific data.

4. Receiving electronics distribute the data within the cabinets.

5. Driver electronics control the LEDs.

6. The pixels emit the visible image.

Enlarge diagram

Scroll across the enlarged diagram to read its labels.

Figure 2. A high-level LED display signal path.

Power distribution, network control, monitoring and redundancy operate alongside the signal path. Product architectures differ, and some functions may be combined into one device or distributed across several devices.

2.5 Power, heat and efficiency

LEDs convert electrical energy into light more efficiently than incandescent sources, but an LED display is not just an array of LEDs. It includes driver electronics, power conversion, control systems and sometimes active cooling. Electrical input that is not emitted as useful light becomes heat and must be managed.

Power demand varies with:

• display area and pixel density;

• selected luminance;

• image content and average picture level;

• duty cycle and operating hours;

• driver and power-supply efficiency;

• ambient temperature and cooling strategy.

A peak or maximum power figure is not the same as typical energy use. Procurement and electrical planning should ask how each figure was measured and whether it represents maximum white, a defined test pattern or expected content.

2.6 Redundancy and monitoring

Displays used for broadcast, transport, live events or critical information may include redundant signal paths, power arrangements or network control. Monitoring can report temperatures, power states, data status and other indicators. These features improve resilience only when the design removes meaningful single points of failure and when alerts are linked to an operational response.

Redundancy should therefore be documented as an architecture, not accepted as a label. The questions are: which elements are duplicated, how failover is triggered, what remains a single point of failure, and how the organisation verifies that the fallback path is available.