DeltaE 94 indicates the variation of hues. This definition enhances the original CIE Lab formula from 1976 by corrective factors considering the saturation of the colour samples. At this, we are using factors that are common in the graphic industry. In terms of neutral gray colour samples, both formulas would deliver an identical result. The more saturated colour samples are the lower the colour distance is in DeltaE 94 when compared to the original definition. Nevertheless, even the newer formula does not reach full equidistance.
That is what makes a general referral for the recommended range of deviation so difficult. Apart from very few saturated colour samples, not even a direct comparison reveals any colour difference when DeltaE equals one.
If DeltaE is higher than five one normally has to speak of a large deviation. In our procedure, we gauge several hues (primary and secondary colours, as well as some tertiary colours), as secondary and particularly tertiary colours do supply good hints as to the linearity of a display.
In comparison to sRGB, displays with enhanced colour space naturally show considerable deviations from the sRGB reference. It is desirable if the factory setting provides a most neutral gray axis. For the white point, we specify the deviation from DeltaE to our target white point in the chart that is prefixed each time. In a non-calibrated state, it may well indeed come to larger deviations, which is unproblematic in the first instance in most cases: As recommendation, the target value is seldom appropriate. Right now, the distance between white point and black body curve should be as low as possible, which can be verified with the second value. This one should not exceed a DeltaE of three. After calibration, both values have to be considerably lower. Our target white points (D50 and D65) only show a minor shift from the black body curve.