by clicking the arrows at the side of the page, or by using the toolbar.
by clicking anywhere on the page.
by dragging the page around when zoomed in.
by clicking anywhere on the page when zoomed in.
web sites or send emails by clicking on hyperlinks.
Email this page to a friend
Search this issue
Index - jump to page or section
Archive - view past issues
FLEXO Magazine : June 2014
matically. Color difference is a better indicator of reproduction quality than optical density. The dramatic color change arises because color difference is a positive only distance metric. If it were possible to have negative color differences, then the difference line would ex- tend downward and produce negative values for film weights less than the aim value of 1.0-μm. Still, the slope of the color difference line is much steeper. Density, on the other hand, allows the press- man to maintain the amount of ink that is transferred to substrate with reasonable latitude and a very nearly linear relationship to his ink key indicator variables. In Figure 4, in the panel for the yellow ink, a 20 percent range of film weight (0.9 -μm. to 1.1-μm.) results in about a 30 point change density (±15 points). If we plot the logarithm of the tristimulus values (X, Y, Z), scaled from zero to 1.0 instead of zero to 100, against the equivalent logarithm of the spectral product, we obtain curves like that seen in Image 5, computed from some early GRACoL characterization data. Image 5 is a set of very informative diagrams. The yellow ink mapping is very consistent, no matter which spectral product, CIE Z or ISO TB is applied. But both the cyan and magenta have a large number of stray or inconsistent mappings. This is because the visual system is not opti- mized to see the absorption bands of process printing inks. In mathe- matical terms, the ISO spectral products are said to be “sharper” than the visual spectral products. Fortunately, this is not a new discovery and those giants who have gone before us in digital imaging have devel- oped a number of different versions of sharpened tristimulus values. One simple set was reported by Viggiano in a paper whose aim was to simulate the Status T density values with tristimulus values. He created a set of sharpened sensor values, which he named R, G and B. They are related to X, Y and Z by a simple matrix transform as shown: When we “sharpen” the XYZ values into RGB and plot the sharpened colorimetric density against the ISO Status T density, we get much better re- sults. Since ISO Visual density is based on CIE Vλ, which is also identical to the CIE Y tristim- ulus value, the black spectral product is already identical to the colori- metric spectral product. Image 6 shows much improved correlation between colorimetric based density and ISO Status T density. We would see about the same level of disagreement if we plotted ISO Status A or Status M against Status T and either of those methods can control the color of process primaries, as well as Status T. In fact, the B spectral product in Status E is about as different from T as T is from Z. Thus any graphic reproduction property that might be tracked using Image 4: Curves are showing ΔE* (blue curve) and Status T Density (green curve) versus film thickness. Image 5: Correlation of ISO Status Density and Colorimetric Density 42 FLEXO | JUNE 2014