Short answer
Because each system measures different things for different purposes. RGB counts light for a display; CMYK counts ink for a press; HSL and HSV rearrange RGB geometry; XYZ predicts matches; CIELAB and OKLab model perceived differences; Munsell, NCS, RAL and Pantone point to physical references. Even words they share — hue, lightness, saturation — are defined differently, so one colour's 'hue' can be 210° in HSL, 279° in CIE LCh and 253° in OKLCH, and all three are correct.
The table on this page takes a single bright mid-blue, sRGB #1E90FF, and expresses it in every model Colourwise computes. Nothing about the colour changes from row to row; only the question being answered does. RGB and HEX say how hard to drive a display's three primaries. HSL and HSV rearrange those three numbers into an angle and two ratios. CMYK gives a naive ink recipe. XYZ gives tristimulus values that predict a match for the standard observer. CIELAB, LCh, OKLab and OKLCH place the colour in spaces designed so that distances roughly track perceived differences. Each is a different coordinate system over the same object.
Most confusion comes from shared vocabulary. HSL 'lightness' is the average of the largest and smallest RGB channels, so this blue scores 56%; HSV 'value' is the largest channel, so it scores 100%; CIELAB L* models perceived lightness and gives about 59; OKLab L gives 0.65 on a 0–1 scale. 'Hue' is an angle in every polar model, but each starts from a different zero and spaces hues differently: HSL puts this colour at about 210°, CIE LCh at about 279° and OKLCH at about 253°. 'Saturation' in HSL, 'chroma' in LCh and OKLCH, and 'chromaticness' in NCS are three different measures of colourfulness. Numbers can never be copied between models, only converted.
A useful habit: whenever you write a colour number, write its model and, for CIELAB, its reference white. '65, 0.19, 253' means nothing; 'oklch(0.65 0.19 253)' is a colour.
Even within one model, results depend on settings that are rarely shown. CIELAB values depend on the reference white (D50 or D65) and, for measured surfaces, on the illuminant and observer. RGB-derived values depend on which RGB space the input was in: the same HEX interpreted as Display P3 gives different Lab numbers. CMYK depends entirely on the printing condition. Rounding matters for round trips: converting to 8-bit RGB and back can move values by a little. And systems built on physical samples — Munsell, NCS, RAL, Pantone — only map to numbers through measured data under stated conditions, which is why two websites can give two different HEX values for the same catalogue colour.
The deeper disagreement is about distance. Two pairs of colours that are equally far apart in RGB can look very differently far apart; CIELAB distances are better but still uneven, especially in blues and saturated colours; OKLab and CIEDE2000 are better again, but no formula perfectly predicts every observer. This is why 'how close is this colour?' depends on which model answers it, and why tools that find the nearest named colour or paint can reasonably return different answers. Colourwise states which model and formula it uses wherever it reports a difference.
| Model | Values | What the numbers measure |
|---|---|---|
| HEX | #1e90ff | sRGB code values in base 16 |
| RGB | 30, 144, 255 | Encoded sRGB, 0–255 |
| HSL | 209.6°, 100.0%, 55.9% | Hue from RGB hexagon; 'lightness' is (max+min)/2 |
| HSV | 209.6°, 88.2%, 100.0% | Same hue as HSL; 'value' is the largest channel |
| CMYK (naive) | 88, 44, 0, 0 | Arithmetic inversion; not a press value |
| CIE XYZ (D65) | 28.6, 27.4, 98.4 | Y is relative luminance ×100 |
| CIELAB (D65) | 59.4, 10.0, -63.4 | L* lightness; a*, b* opponent axes |
| CIE LCh (D65) | 59.4, 64.2, 278.9° | CIELAB in polar form |
| OKLab | 0.652, -0.055, -0.182 | L on a 0–1 scale |
| OKLCH | 0.652, 0.190, 253.2° | OKLab in polar form |
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
Colourwise analysisStrong evidence
The sRGB colour #1E90FF has an HSL hue of about 210°, a CIE LCh (D65) hue of about 279° and an OKLCH hue of about 253°, and a 'lightness' of about 56% in HSL, 59 in CIELAB and 0.65 in OKLab.
Based on: Computed by Colourwise with its conversion engine; see the table on this page.
Source: International Electrotechnical Commission (IEC) webstore and catalogue; CIE 015:2018 Colorimetry, 4th edition; A perceptual color space for image processing (Oklab)
FactStrong evidence
CIELAB values depend on the chosen reference white; CSS lab() uses D50 while many display-oriented tools use D65, so the same colour yields different a* and b* values.
Source: CSS Color Module Level 4; CIE 015:2018 Colorimetry, 4th edition
Colourwise interpretationModerate evidence
Differences between models reflect different purposes — display drive, ink, match prediction, perceived difference, physical reference — rather than errors in any one of them.
Based on: Colourwise synthesis of the design aims stated in the sources for each model on this pillar.
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.