What calibration actually corrects
No two LED packages are identical. Manufacturers sort them into bins by brightness and chromaticity, but a bin still spans a range, and a display built from several batches will carry that spread onto the wall. Calibration compensates for it by adjusting the drive signal per module, or per pixel, so every part of the screen hits the same target instead of its own natural output.
The practical result is a wall that reads as one surface. Without it, a large white field can show patchy blocks, a slight green or magenta cast in one corner, or seams that only become obvious when the content changes. These differences are easiest to see on low-brightness corporate content and skin tones, and hardest to see on busy video, which is exactly why they are often missed during a demo.
Brightness uniformity is not colour uniformity
The two are measured separately and corrected separately. Brightness uniformity describes whether one area of the screen is brighter than another at the same signal level. Colour uniformity describes whether the same white or the same grey has the same hue across the wall. A screen can pass one and fail the other.
This distinction matters when you write acceptance criteria. Naming only brightness leaves a supplier free to deliver a screen that is evenly bright but shifts in colour from cabinet to cabinet, which is the failure most viewers actually notice first. Ask for both, and ask how each was measured.
Why a screen drifts apart after installation
Calibration is a snapshot, not a permanent state. LEDs lose output gradually with operating hours, and the red, green and blue emitters do not age at the same rate. Temperature, drive current and the mix of content all affect how fast that happens, so a display running bright white content for long days will not age the same way as one running darker material.
The consequence is that a wall calibrated in the factory can drift back into visible variation over time. How quickly depends on the installation, the duty cycle and the environment. Plan a repeat calibration once the display has been running. Commissioning should not be the last time uniformity is checked.
Factory calibration is not the final word
Modules are usually corrected before they leave the factory, and that is valuable: it removes the largest batch differences. It is not the same as calibrating the assembled wall in its final position. Cabinets are joined, powered and driven as a system only once they are installed, and the control system, the power distribution and the physical alignment all enter the picture at that point.
Ask whether commissioning includes an on-site calibration pass, who performs it, and what equipment is used. A handheld meter held at a few positions is not the same exercise as a planned measurement across the whole surface. Also ask whether the display is given warm-up time before measurement, because readings taken on a cold panel will not match readings taken after it has stabilised.
Replacement modules and colour matching
The hardest uniformity problem usually arrives later, when a damaged module is swapped out. A replacement from a different production run will carry its own brightness and colour characteristics, and on a fine pitch wall even a single fresh module can stand out against its aged neighbours.
Plan for this before you need it. Keeping spare modules from the same batch as the original build is the simplest mitigation. Where that is not possible, expect to recalibrate the affected area after replacement, and ask in advance whether the control system can store and reapply per-module correction data after a swap.
Writing acceptance criteria you can actually check
Uniformity requirements are only enforceable if they come with measurement conditions. A useful clause names the target, the tolerance, the device used to measure, the brightness level the screen is set to, the warm-up period, the ambient light conditions, and how many points across the surface are sampled. Without those, a figure on a datasheet cannot be verified on site.
Also state what happens if the screen misses the target: whether the supplier recalibrates, replaces modules, or accepts a documented deviation. And keep the record. A before-and-after measurement report gives you a baseline for later comparison when you suspect the wall has drifted.
What to confirm before you sign off
Four questions cover most of the risk. Is calibration performed per pixel or per module? Is there an on-site pass after installation, or only a factory pass? What are the measurement conditions, and will you receive the report? Finally, what is the process and cost basis for recalibration after a module replacement or after a period of operation?
Uniformity is also bounded by physics. Calibration cannot rescue a cabinet that is mechanically misaligned, a panel with viewing angle mismatch, or optical differences between surface finishes. Where a wall will be viewed off-axis, that limitation is worth understanding before the acceptance criteria are drafted.
A wall that drifted after install
Factory calibration looked even on day one, but after run-in a few cabinets pulled colour. A planned second pass closed the gap; without the record, matching would have meant recalibrating the whole wall.
A swapped module that stood out
A failed module was replaced with a fresh one at factory default. Because the calibration file was kept, the new module was matched to the wall instead of looking like a patch.
Putting this into practice: calibration and colour uniformity
- Agree brightness and colour targets separately, in writing Name the measurement conditions, brightness level, colour temperature and dark-room state, so the result is something you can actually check at acceptance.
- Treat factory calibration as a start, not the final word Plan a repeat pass after the wall has run in, when early drift shows. The factory result is the starting point, not the finish.
- Plan for replacement modules and a matching record Keep the calibration file so a swapped module can be matched to the wall, not just to the factory default. A missing file forces a full recalibration later.
Assumptions to check on calibration and colour uniformity
“Factory calibration is the final word.” — LEDs drift after run-in; a second pass is normally needed. The factory result is the starting point, not the finish.
“Calibrating to a uniform target reduces brightness.” — Uniformity work adjusts relative output across modules. Done well it preserves the target brightness while removing patches.
“Brightness uniformity is the same as colour uniformity.” — A wall can be evenly bright yet show colour patches, or evenly coloured yet unevenly bright. They are separate measurements.
What to carry forward on calibration and colour uniformity
Treat calibration as a process with a start, a record and a follow-up, not a line on a specification sheet. Agree brightness and colour targets separately, name the measurement conditions in writing, and plan for a repeat pass after the display has run in. If you are specifying a wall where uniformity genuinely matters, share your viewing conditions and content type on the contact page before the order is placed.