Quick Answer
An LED display turns a video signal into light in four steps: a video processor scales and splits the incoming frame, a sending card distributes it to receiving cards, each receiving card hands its portion to driver ICs, and the drivers switch individual red, green and blue LEDs on and off thousands of times per second using pulse-width modulation. Brightness comes from how long each LED stays lit within a cycle, not from how brightly it burns — which is why refresh rate, greyscale depth and driver quality determine what the eye actually sees.
Key Takeaways
- LEDs are switched, not dimmed. Brightness is created by pulse-width modulation — turning the LED on for a longer or shorter fraction of each cycle.
- Colour comes from three emitters. Red, green and blue lamps at different intensities combine into the full range the eye perceives.
- Greyscale depth sets the smoothness of gradients. It is a separate specification from refresh rate, and both appear on serious spec sheets.
- The image is cut into cabinets, then modules, then pixels. Each level of the chain imposes its own limits on resolution and reliability.
- Colour accuracy depends on the whole chain, from LED bin selection through driver linearity to calibration coefficients stored in the cabinet.
What Happens Between a Video Signal and a Lit Pixel?
The path from source to light passes through a fixed sequence of components, and knowing the sequence makes it much easier to understand where problems occur.
Step 1: video processing
A video processor receives the source — HDMI, SDI, DisplayPort or a network stream — and performs three jobs. It scales the incoming frame to the display's native pixel dimensions, it converts colour space and frame rate as needed, and it maps the canvas across the physical cabinet layout. For a non-rectangular or unusually proportioned wall, this mapping is where the screen's geometry is defined. Poor scaling here is visible as softness before the signal ever reaches the LEDs.
Step 2: distribution by the sending card
The sending card takes the processed frame and splits it into streams, one per chain of cabinets, then transmits them over Cat5e/6, fibre or wireless links. This is where the practical limit on screen size originates: each output port supports a maximum pixel load, so a large wall needs multiple ports or multiple sending cards.
Step 3: decoding at the receiving card
Each cabinet contains a receiving card that accepts its stream, decodes it, and distributes the pixel data to the driver ICs on each module. The receiving card also holds the cabinet's calibration coefficients — the correction values that make this cabinet match its neighbours.
Step 4: switching by the driver IC
Driver ICs are the final stage. Each drives a fixed number of output channels, switching LEDs on and off at high frequency. The driver's characteristics set the maximum refresh rate and greyscale depth the panel can achieve, and its linearity affects how evenly adjacent pixels respond to the same input value.
How Does an LED Produce Colour?
Each pixel is not a single device. It is a group of three LED lamps — one red, one green, one blue — placed close enough together that the eye merges them at normal viewing distance. By varying the on-time of each of the three lamps independently, the display produces any colour within the triangle those three primaries can define.
Two consequences follow. First, colour gamut is limited by the lamps. A display using wider-gamut green and red emitters can reproduce more saturated colours; this is why LED brand and wavelength bin appear in serious specifications. Second, white is not a colour the display has — it is produced by running all three lamps together, and the accuracy of that white depends on balancing the three primaries across the entire screen.
This is also where bin selection enters. LEDs are manufactured on wafers and sorted into bins by brightness and dominant wavelength. Panels built from tightly binned lamps need less correction and hold uniformity better over time. Panels built from wide bins can look uneven in solid colours even when every other specification matches.
How Does the Display Create the Illusion of Motion?
The mechanism is pulse-width modulation. Rather than running an LED at half current to produce half brightness — which would shift its colour and reduce efficiency — the driver switches it fully on and fully off within each refresh cycle. A 50% duty cycle looks like half brightness to the eye because the switching happens far faster than human persistence of vision.
This explains two specifications that buyers frequently confuse.
| Specification | What it controls | Why it matters |
|---|---|---|
| Refresh rate (Hz) | How often each LED is re-switched per second | Determines whether cameras capture the image cleanly |
| Greyscale depth (bit) | How many discrete brightness steps fit inside each cycle | Determines how smoothly gradients and shadows render |
| Scan mode | How many rows of LEDs share one driver channel | Affects brightness, refresh ceiling and cost |
Greyscale depth is the subtle one. A 14-bit driver can express 16,384 brightness levels per colour channel; a 10-bit driver only 1,024. On content with smooth gradients — dusk skies, skin tones, slow fades — the difference is visible as banding on the lower-depth panel, even when refresh rate and pitch are identical.
Scan mode explains why two panels with the same pitch can differ in brightness. A 1/8-scan panel lights one row at a time across eight rows; a 1/32-scan panel divides the work differently. Lower scan ratios generally allow higher brightness and refresh but require more driver channels, which raises cost.
How Is the Image Divided Across Cabinets?
A large LED display is assembled from cabinets, and each cabinet from modules. This hierarchy is what allows unlimited screen sizes and module-level repair — from a single-column LED poster display to a wall of several hundred cabinets — but it also introduces the alignment question that determines whether the wall looks seamless.
Each cabinet contains several modules, typically 250 × 250 mm or 320 × 160 mm, held in a frame with a specified flatness tolerance. When cabinets are joined, small differences in flatness and in module brightness become visible as grid lines — particularly when the screen is viewed at an angle or shown in a solid colour.
Three practices reduce this. Die-cast aluminium cabinets maintain their geometry through thermal cycling better than steel or plastic. Tight flatness tolerances, quoted in fractions of a millimetre, are what you are paying for in a premium cabinet. Per-cabinet calibration corrects the brightness and colour differences that remain. A screen that is flat and calibrated looks continuous; one that is either but not both shows seams.
Why Does Refresh Rate Change What a Camera Sees?
A camera does not see continuously. It samples the scene during an exposure window, and if the display is mid-cycle when the sample is taken, the camera captures a fraction of the LEDs in their "off" state. Across a screen, this appears as horizontal banding, and because the display and camera run on independent clocks, the bands roll upward or downward through the image.
The fix is to raise the refresh rate far above the camera's frame rate — a factor of roughly sixty times the frame rate is comfortable. For a camera shooting at 50 or 60 frames per second, that puts the practical floor at around 3,000 Hz, which is why 3,840 Hz has become the de facto standard for anything that will be filmed.
Two further points matter in practice. Shutter speed interacts with refresh rate: a camera with a very short exposure can capture artefacts even at high refresh rates, so broadcast work sometimes specifies 7,680 Hz. And grey scale is sometimes sacrificed to reach a refresh figure — a quoted 3,840 Hz at reduced bit depth is not equivalent to 3,840 Hz at full depth, so ask for both numbers together.
What Limits the Maximum Size of a Single Display?
LED walls scale further than almost any other display technology, but not without limits. Four constraints define the practical ceiling.
| Constraint | What it limits | How it is solved |
|---|---|---|
| Sending card output capacity | Pixels per output port | Add ports or additional sending cards |
| Signal transmission distance | Metres from source to first cabinet | Fibre links, or distributed processors |
| Power distribution | Total load per circuit | Multiple circuits, distributed supplies |
| Structural load | Weight per square metre | Steel sub-frame, rigging design |
In practice, the constraint that surprises buyers is power distribution. A large high-brightness wall draws substantial current, and the available supply in the building — not the display's capability — is often what determines the maximum size. This is why power planning belongs in the earliest design conversation rather than in the installation phase.
How Does Calibration Fit Into the Chain?
Calibration is the final step that turns a collection of cabinets into a single image. It works at the cabinet and module level, correcting the two variables that the manufacturing process cannot hold perfectly: brightness uniformity and colour consistency.
The process measures each cabinet with a colourimeter, computes correction coefficients, and stores them in the receiving card. On playback, the card applies those coefficients to every incoming value, so a module that runs 3% bright is quietly reduced to match its neighbours. The viewer sees a uniform surface; the display is actually holding each module at a slightly different output.
Three practical implications for buyers:
- Ask whether calibration is done per module or per cabinet. Module-level calibration corrects more variation and is the standard for premium installations.
- Request that correction data ships with the order. If a receiving card is replaced in year three, the saved coefficients allow the new card to be restored to the same state.
- Plan for re-calibration. LEDs age at slightly different rates, so uniformity drifts over years. Periodic re-calibration — typically every two to three years in demanding installations — restores the original appearance at a small fraction of the original cost.
This is also the reason two displays with identical LED and driver specifications can look different in the field: one was calibrated against a documented report, and the other was not. The principle scales across applications — from a stadium wall down to the compact panels in our digital signage range.
Frequently Asked Questions
Does an LED display use one LED per pixel?
No. Each pixel is a cluster of three LEDs — one red, one green and one blue — close enough together that the eye merges them. Colour is produced by varying the relative on-time of the three lamps. This is why pixel pitch refers to the spacing between clusters, and why colour accuracy depends on how consistently the three primaries are matched across the panel.
Why is refresh rate measured in Hz when the eye cannot see flicker above 1,000 Hz?
Because the specification exists for cameras, not eyes. A camera samples the scene during a fixed exposure window, and if the display is mid-cycle it captures partially lit LEDs as banding. Raising the refresh rate far above the camera's frame rate ensures every sample catches complete cycles. If your screen will never be filmed, a lower refresh rate is acceptable and cheaper.
What is the difference between greyscale depth and refresh rate?
Refresh rate is how often each LED is re-switched per second; greyscale depth is how many distinct brightness levels fit within each switching cycle. Refresh rate determines whether cameras see banding; greyscale depth determines whether gradients look smooth. A panel can score well on one and poorly on the other, which is why both figures should appear together on a specification sheet.
How does an LED display make white?
By running all three colour lamps together and balancing them so the result appears neutral. Because white is a combination rather than a dedicated emitter, its accuracy depends on how well the red, green and blue lamps are matched and calibrated across the whole screen. This is one of the reasons wide-bin panels show colour shifts that tight-bin panels do not.
Why do LED walls sometimes show visible seams?
Two causes, usually together: cabinet flatness tolerance and module-to-module brightness differences. When cabinets are joined with slight variations in depth, the seam catches light at an angle. When modules differ in brightness, a grid pattern appears in solid colours. Die-cast cabinets with tight tolerances, plus per-module calibration, address both. Ask for the flatness figure in millimetres before ordering.
Can an LED display be repaired pixel by pixel?
On SMD panels, an individual faulty lamp can be replaced by a technician with a hot-air rework station, though in practice modules are usually swapped and repaired later off-site. On COB panels the LEDs are bonded under a protective layer, so repair happens at module level by design. This is the main repairability trade-off between the two packaging types — see our LED rental panel range for modular cabinets designed around fast module replacement.
About the Author
Written by the VISLAT engineering team. VISLAT (Shenzhen Vision Latitude Co., Ltd.) manufactures commercial LED displays — LED poster displays, LED DJ booths, rental panels, outdoor TVs and digital signage — and has supplied projects in more than 70 countries. All products carry CE, FCC and RoHS certification and a 3-year warranty with 24/7 engineering support.