The applicability of modern color differencing equations in the Graphic Arts industry
At TAGA 2007 conference in Pittsburgh I presented a paper about color-differencing equations and the human eye. The color test was done with inexperienced observers. One of the main results was that the DE2000 equations correspond quite well with the perceived color differences, followed closely the DECMC2:1 equation.
A very similar test was carried out with experienced color observers at the IPA technical conference in Chicago in June 2008 and local print professionals. Besides having to judge a solid color chip, they were also asked to rank visual differences in four test images.
As an addition to the test carried out in 2007, the DIN99 equation was also tested to see if it relates better to the perceived color differences. Interestingly the DE2000 equation related the best with how the observers perceived the color differences of the solid color patches.
The results from the test carried out in 2007 were revisited to see how well they relate to the DIN99 equation. The DIN99 color differencing equation did not correlate better with the data gathered from inexperienced observers.
Experienced observers were also asked to rank color differences of solid color patches. The perceived differences related well with the numerical differences given by the DE2000 equation. Based on the experiments carried out in 2007 and 2008 it can be said that the DE2000 equation should be used more widely in industry since it relates well with the way experienced and inexperienced observers perceive color differences.
Another question is how do any of the modern color-differencing equations relate to differences present between two similar images displayed on a computer screen? Due to the large number of new soft proofing systems being introduced into the graphic arts industry, the test images were shown on a color-calibrated display.
Four ISO SCID images were chosen and modified in regards to lightness, chroma and hue. The images that were chosen varied in chroma from very colorful to mostly neutral colors. In the neutral colors all observers identified differences quite easily, but in the other images many observers had difficulties identifying slight changes in regards to changes in lightness, but could readily identify differences in chroma and hue. The observers more readily reported chroma and hue differences than differences in lightness. The most sensitive image for small color changes contained large neutral gray areas, as was to be expected.
Threshold values in relation to differences in Lightness, Chroma and Hue will be presented and also which of the newer colour differencing equations relates better to the perceived color differences.
