Salt
#efede6
A neutral that has no useful hue — it reads as a grey. It is very light — it works as a page background, not as text.
Hex
#efede6
RGB
239, 237, 230
OKLCh
94.6% 0.010 93.6°
CIELAB (D50)
93.8, -0.2, 3.6
What is true of this colour, and of no other
- 118 other named colours sit within ΔE00 2 of it — this part of the space is crowded, and the names here are finer distinctions than most people make.
- It uses only 12% of the chroma available at this lightness and hue — there is a far more vivid version of this colour inside sRGB.
- As text on white it measures APCA Lc 8.2; on black, Lc -96.1. The sign is the polarity, which the WCAG ratio discards.
- Built into an eleven-step scale, 10 of the 10 adjacent pairs stay ΔE00 2 apart — see the ramp.
- All three dichromacies see it as very nearly the same colour you do: its hue is not the part carrying its identity.
Text on this colour
White text
1.17:1 — fails AA
Black text
17.93:1 — passes AA
Black text is the better of the two at17.93:1. That is enough for body text.
As text on white, it needs to change
#787771 is the nearest colour of the same hue that passes AA on white, at 4.50:1. Only the lightness moved.
The same colour, with one cone missing
As you see it
#efede6
protanopia
#edede6 · ΔE00 1.1
deuteranopia
#eeeee6 · ΔE00 1.1
tritanopia
#eeeee0 · ΔE00 3.4
The twelve nearest named colours
Ranked by CIEDE2000 — perceptual distance, not hex arithmetic. The closest isFuji Snow at ΔE00 0.422: below 1.0 — an ordinary observer cannot tell these apart, even side by side. What the numbers mean.
- Fuji SnowΔE00 0.422 · #f1efe8
- White FurΔE00 0.603 · #f1efe7
- CarraraΔE00 0.644 · #eeebe4
- Commercial WhiteΔE00 0.675 · #edece6
- Melting IciclesΔE00 0.702 · #ecebe4
- Broken WhiteΔE00 0.732 · #eeebe3
- Cotton FieldΔE00 0.76 · #f2f0e8
- Hacienda WhiteΔE00 0.765 · #f0ede7
- Gin TonicΔE00 0.777 · #ecebe5
- Nano WhiteΔE00 0.793 · #f2f0ea
- Cloud DancerΔE00 0.81 · #f1eee8
- Traffic WhiteΔE00 0.871 · #f1f0ea
9 of these have no page of their own: pages exist for the 4,961 colours the dataset marks as good names.