ToonTone Proofing Lab

Color QC · CIEDE2000 matching · 5 rounds

Play ToonTone

What Is Delta E (ΔE)?

Delta E is the number that answers the question every printer, designer, and paint matcher eventually asks: how different are these two colors, really? Not "are they different" — your eye can answer that some of the time — but by how much, on a scale where a 2 is a 2 whether you're comparing two navy inks or two skin tones. ΔE is the unit of that scale: the measured distance between two colors in a space designed to mirror human vision.

CIELAB: The Map Delta E Measures On

To measure a distance you need a map, and the map ΔE uses is CIELAB (also written L*a*b*), published by the CIE in 1976. It describes any color with three coordinates:

  • L* — lightness, from 0 (black) to 100 (white)
  • a* — the green–red axis: negative values lean green, positive lean red
  • b* — the blue–yellow axis: negative values lean blue, positive lean yellow

The point of CIELAB is perceptual uniformity: it was built so that one unit of distance anywhere on the map should look like roughly the same amount of difference to a human observer. That's what RGB can't give you — in RGB, a jump of 10 units might be invisible in one region and glaring in another, because RGB describes how much light a screen emits, not how a person sees. Converting a color to CIELAB is like converting raw pixels into the eye's own bookkeeping. Once two colors sit on this map, the difference between them becomes geometry.

Why CIEDE2000 Is Fairer Than ΔE76

The original formula, ΔE76, is just straight-line Euclidean distance in CIELAB: simple, fast, and honestly not bad. The problem is that CIELAB turned out to be only approximately uniform. Careful testing showed systematic mismatches:

  • The eye tolerates bigger chroma differences in vivid colors. Two saturated reds can be further apart than two grays before anyone notices. ΔE76 treats both gaps identically; CIEDE2000 adds a chroma weighting term that forgives distance in high-chroma regions and tightens tolerance near neutrals.
  • The eye is stricter about lightness in mid-tones. CIEDE2000's lightness weighting reflects that sensitivity curve.
  • Blues misbehave. Around the blue region, hue differences interact with chroma in a way a plain distance can't capture, so CIEDE2000 adds a rotation term (RT) that twists the calculation there.

CIEDE2000 (ΔE00) keeps the same three-part skeleton — lightness, chroma, hue — but wraps each in a weighting function derived from real perceptual experiments. The result: its numbers track trained human judgement far more consistently. When a proofreader and a press operator disagree about whether a job matches, ΔE00 is the arbiter the industry reaches for. It's what this lab uses, with the standard reference conditions (D65 illuminant, kL = kC = kH = 1).

Delta E Values in Practice

ΔE00What it means
< 1Not perceptible — two proofs read as identical
1–2Visible only to a trained eye, usually side by side
2–10A clear, nameable difference at a glance
> 10Different colors, full stop

Context shifts the stakes. A packaging printer may hold brand colors to ΔE < 2 across a run; a textile buyer might accept 3; a billboard nobody compares to a reference can drift further. But the scale itself is stable: a ΔE of 1.5 means the same class of "barely there" in any shop.

How ToonTone Scores Your Proofs

Here is the exact math this game runs, because a lab that won't show its formula shouldn't be trusted:

  • Your swatch and the target swatch are converted from RGB to CIELAB (D65 reference white).
  • The full CIEDE2000 formula computes ΔE between them — weighting terms, blue-region rotation, and all. No shortcuts, no ΔE76.
  • score = 100 − ΔE, clamped to the 0–100 range and rounded to a whole number. A dead-on match (ΔE under 0.5 after rounding) scores 100; anything ΔE ≥ 100 scores 0, though such misses are rare outside of submitting the wrong hue entirely.
  • Five rounds make a game: 500 possible points. Your average sets the press rating — 60+ Color Apprentice, 75+ Color Pro, 90+ Color Master.

Two consequences worth knowing before you blame the referee. First, because CIEDE2000 tightens tolerances in low-chroma regions, pastel targets punish small hue errors more than saturated targets do — the formula is mirroring your own visual system. Second, the scoring is purely numerical: your display, room light, and fatigue don't enter the calculation. They only affect what your eye reported. Ready to be measured? Try the ΔE game and see your own numbers, or learn how to play ToonTone step by step. The sliders themselves are covered in the HSB color model guide, and everything else is in the FAQ.

FAQ

What is a good Delta E value?

In print proofing, ΔE00 under 1 is treated as a match, under 2 as commercially acceptable for most brand work. In this game, an average under 10 across five rounds (a Color Master rating) means your eye is doing pressroom-grade work. Anything under 25 average is a respectable shift at the proofing table.

Can Delta E be zero?

Yes — it means the two colors are numerically identical in CIELAB. In the game it happens when you land on the exact target RGB values. It is rare and worth a small celebration: ΔE 0 with sliders is threading a needle.

What's the difference between ΔE76, ΔE94, and CIEDE2000?

All three measure distance in CIELAB-type spaces. ΔE76 is plain Euclidean distance. ΔE94 added weighting for chroma and lightness, tuned on graphic-arts data. CIEDE2000 refined that with better lightness, chroma, and hue weightings plus a rotation term for blue hues, and it agrees with human judgement across the widest range of colors. ToonTone uses CIEDE2000.

Does Delta E depend on lighting?

The CIELAB coordinates behind ΔE are computed relative to a reference illuminant — this game uses D65, standard daylight. Real-world viewing still matters: a print checked under a warm desk lamp won't match a ΔE report computed for daylight. That's why pressrooms use standardized viewing booths.

Why does the same ΔE feel bigger on some colors than others?

Because no formula is perfectly uniform — CIEDE2000 is the closest industry standard, but human vision still has local quirks. Treat ΔE as a very good ruler, not an oracle: it tells you how much difference there is, and your eye confirms whether anyone will care.