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 : February 2009
TECHNOLOGIES & TECHNIQUES The spectral component of the illumination is another source of error, especially if the substrate or the ink fluoresces. If two color measurement devices have a different amount of UV in their illumination, then the activation of the fluorescence will be different, and the devices will read differently. This is a particular problem for measurements made on a web. Online systems typically use either xenon strobes or LEDs for illumination. The strobes have an appreciable amount of UV, the LEDs have virtually none. Handheld spectrophotometers, how- ever, generally use incandescent lighting, which has some UV, but not as much as a strobe. Scattered light poses yet another problem. Light scatters within the camera and is measured at the wrong position in the image. Figure 3 shows light rays (in green) that are imaged on the detec- tor in the correct place, and light rays (in red) that are imaged in the wrong place. On the left, light scatters when it reflects from the lens holder. On the right, light that should be collected at one spot on the detector, but instead reflects from the surface of the detector and then reflects again from the protective glass cover plate. Thus, light from a very light portion of the web may con- taminate a very dark patch to be measured. Scattered light can increase the apparent reflectance of an area by a few percentage points. While this may not sound like much, it can have considerable effect, especially in the darker portions of an image. The effect of adding 1 percent reflected light to a black area with reflectance of 1 percent will change the L * value from 9 to 15.5. Correction for this scattered light is possible, but difficult be- cause it depends to a varying degree on every pixel in the image. Correction requires a painstaking characterization of the camera. FIGURE 3. Some examples of how light can scatter in a camera. A dirty lens is one particularly nasty cause of scattered light. A dirty lens will add an overall haze to the image, much like view- ing the web through a fog. Pressurized air can reduce the accu- mulation of the dust and mist, but accurate measurements still require periodic careful cleaning of the first glass surface. Even if these engineering issues are properly dealt with, there is still a fundamental limit to the accuracy of colorimetric mea- surements derived from an RGB camera: the spectral response. RGB cameras, unfortunately, do not respond to different wave- lengths of light in the same way that a colorimeter (or the human eye) does. Two objects that have the same CIELAB values may look considerably different to an RGB camera. There have been numerous papers written about various meth- ods to convert RGB values from a camera or scanner into CIELAB values. I found eight papers that provided enough explanation of the methods and experimental data to allow comparison. Generally speaking, the average color error that is reported is from 4 to 10AE. Only three papers report an accuracy that is at or below the absolute minimum requirement of 2.4AE. Two of those papers were reporting errors strictly due to spectral response, so the actual error will be larger than this. The third paper reports accuracy as good as l.2AE, but the measurements are limited to a small number of patches on a single newspaper stock. Based on this, it is doubtful that claims of accuracy at or below 1.0AE are real. TEST OF COLOR DIFFERENCE To meet the ISO "color OK" requirement, one needs to match color against a number that another instrument read from the proof, so one needs accuracy of CIELAB values. But for the pro- Detector Light belongs here. .. Detector Lens holder Web - Protective cover Lens I FEBRUARY 2009 www.flexography.org FLEXO
Sustainable Winter 2009