Delta E Verification: How to Read Your DisplayCAL Report

After DisplayCAL builds a profile, the verification report tells you whether the calibration actually worked. If you have not generated a report yet, the quickstart guide walks you through a full calibration. The headline number is Delta E — here is how to read it like a pro.

Delta E (CIEDE2000, written ΔE₀₀) is a single number that expresses how different a measured color is from its target color, weighted to match human vision. DisplayCAL’s verification report lists the average and maximum Delta E across a set of test patches, plus gamut coverage charts, so you can judge at a glance whether a profile is accurate.

How to interpret Delta E numbers

Delta E is scaled so that the numbers map roughly to what you can see. Use these widely-used thresholds when reading any DisplayCAL report:

Delta E (CIEDE2000)	What it means	Verdict
< 1.0	Difference is imperceptible, even side by side	Excellent — reference-grade
1.0 – 2.0	Barely perceptible; visible only in direct side-by-side comparison	Good — fine for photo and design work
2.0 – 3.0	Noticeable if you look for it	Acceptable for general use
> 3.0	Clearly visible difference	Poor — re-profile and check your setup

How to run a verification in DisplayCAL

1. Keep everything identical to the profiling run. Same display mode (SDR vs HDR — a mode mismatch is the #1 cause of terrible verification numbers, see our HDR guide), same brightness, same ambient light.
2. Open the verification workflow in DisplayCAL and select a test chart — a few dozen to a few hundred patches. More patches give a more honest report; a 20-patch check can hide problems.
3. Measure. DisplayCAL displays each patch and records what the instrument sees, then compares it against the profile’s predictions.
4. Read the report: check average and maximum Delta E first, then scan the per-patch table for the worst offenders, then look at the gamut plots.

How to read the gamut charts

A DisplayCAL verification report usually includes a chromaticity diagram (the familiar horseshoe) with:

• Your display’s measured gamut — a polygon showing the actual colors your panel reproduced.
• Reference gamuts (sRGB, Adobe RGB, DCI-P3) overlaid for comparison, so you can see at a glance which standards your display covers.
• Coverage percentages — e.g. “98% of sRGB”. Coverage tells you how much of a standard fits inside your display’s gamut; it does not tell you how accurate those colors are. A display can cover 100% of sRGB and still be inaccurate — that is what the Delta E numbers are for.

Read the two together: gamut coverage describes the size of the canvas; Delta E describes how faithfully colors are painted on it. You need both numbers to judge a calibration.

Average, maximum, and what “good” looks like in practice

Most DisplayCAL verification reports summarize with two numbers: the average Delta E across all patches and the maximum Delta E of the worst patch. In practice, treat them as answering different questions. The average answers “how accurate is this profile overall” — for photo editing, retouching, and design work, an average under 2.0 is a comfortable result and under 1.0 is excellent. The maximum answers “where does this profile break down” — a maximum above 5 or 6 usually points at specific saturated colors the display cannot reproduce, or at a setup mistake worth investigating rather than accepting.

Do not chase a perfect report. Diminishing returns set in fast: the visible difference between an average Delta E of 1.5 and 0.8 is essentially nil in real images. Time is better spent re-verifying every few months (backlights age, panels drift) than squeezing the last tenth of a Delta E out of a single session. A practical rhythm is to verify after every calibration, and to re-profile when the average drifts above 2 or when you change anything in the signal chain — new GPU driver, new cable, new OS color settings.

When the numbers are bad: a checklist

1. Display mode changed? HDR on during profiling but off during verification (or vice versa) invalidates everything.
2. Wrong spectral correction? Especially on wide-gamut and OLED panels — see the OLED guide.
3. Display settings drifted? Brightness, contrast, or color presets changed after profiling.
4. Ambient light changed? Verifying in a bright room after profiling in the dark skews dark-patch readings.
5. Panel warmed up? Measuring a cold panel gives different results than a warm one — give it 20–30 minutes.

New to the whole process? Start with the latest DisplayCAL and ArgyllCMS, then work through calibration before worrying about verification numbers.

Frequently asked questions