Short answer
A grey card is a spectrally flat neutral surface you photograph in the same light as your subject, so that white balance and exposure can be set from something known. 18% reflectance is used because it sits close to the perceptual middle of the lightness scale — about L* 50 in CIELAB — even though it reflects less than a fifth of the light. A colour target such as the 24-patch ColorChecker goes further, letting software build a camera profile from many known colours.
A camera cannot tell from a scene alone how much of a colour belongs to the light and how much to the object. A surface whose colour you already know removes the ambiguity. A good grey card reflects all visible wavelengths nearly equally, so any colour it shows in the photograph is the colour of the light, and the white balance that makes it neutral is the correct one for everything lit the same way. Put the card where the subject is, facing the same way, take one frame with it, and use that frame for the white balance of the whole series.
Perceived lightness is roughly proportional to the cube root of luminance, not to luminance itself. The CIELAB lightness formula captures this: a surface reflecting 18% of the light has an L* of about 49.5, almost exactly halfway between black at 0 and white at 100. So 18% reflectance looks like a middle grey even though it reflects much less than half the light. The table below computes the relationship for several reflectances, with the exposure difference in stops from 18%. Light meters are calibrated around a similar middle-tone assumption, so metering off a mid-grey card gives an exposure that places the card near the middle of the tonal range.
Meter calibrations vary slightly between manufacturers and standards, so metering a grey card is a strong starting point rather than a guarantee.
The ColorChecker Color Rendition Chart was presented in 1976 by C. S. McCamy, H. Marcus and J. G. Davidson of the Macbeth company, in the Journal of Applied Photographic Engineering. Its 24 patches in a 4 × 6 layout include six neutral steps, additive and subtractive primaries, and natural colours such as light and dark skin, blue sky and foliage, chosen to behave like the real objects across light sources. It became the reference target for camera profiling. X-Rite introduced new colour formulations for charts manufactured from November 2014, and many patches differ visibly from the older reference data, so the chart's edition matters when using published values.
Light the target evenly with the same source as the subject, avoid glare by angling it slightly, and expose so the white patch is bright but not clipped in any channel. Use the reference data that matches your chart's generation, or measure your own chart with a spectrophotometer. Replace grey cards and charts that have faded, yellowed or become dirty: a target is only as good as its known colour. For white balance alone, any trustworthy neutral works; for profiling, the colour patches do the work.
| Luminance reflectance | CIELAB L* | sRGB 8-bit value (R=G=B) | Stops from 18% |
|---|---|---|---|
| 3% | 20 | 48 | −2.6 |
| 9% | 36 | 85 | −1 |
| 12% | 41.2 | 97 | −0.6 |
| 18% | 49.5 | 118 | +0 |
| 36% | 66.5 | 162 | +1 |
| 50% | 76.1 | 188 | +1.5 |
| 90% | 96 | 243 | +2.3 |
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
A neutral surface of 18% luminance reflectance has a CIELAB lightness of about L* 49.5, close to the perceptual midpoint between black and white.
Based on: Computed by Colourwise from the CIE 1976 L* formula (L* = 116·(Y/Yn)^(1/3) − 16) with Y/Yn = 0.18; see the table.
FactModerate evidence
The ColorChecker Color Rendition Chart was presented by McCamy, Marcus and Davidson of the Macbeth company in 1976 (J. Appl. Phot. Eng. 2(3):95–99) and has 24 patches in a 4 × 6 array.
Source: The ColorChecker Pages
FactModerate evidence
X-Rite states that ColorChecker Classic and SG charts manufactured after November 2014 use new colorants with revised reference data, so a chart's age matters when using published patch values.
Source: New color specifications for ColorChecker SG and Classic charts; The ColorChecker Pages
Colourwise interpretationModerate evidence
Setting white balance from a neutral target in the same light is more reliable than automatic white balance, because it removes the ambiguity between illuminant and surface colour.
Based on: Follows from auto white balance being ill-posed, as described by the Night Sight engineers, and from white balance being defined as the gains that make greys grey.
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.