Spectrophotometers do not allow the measurement of complex, color-patterned surfaces, since they inevitably record several spectra, while determining the average of them, i.e. "single-color" spectral readings.
In addition, the visual impressions that people perceive often cannot be verified through measurements with a spectrophotometer, since a person's subjective visual impression is greatly affected by influences such as contrasts in color and brightness, which can't be recorded in a spectral measurement.
In this example of different grains, the measurement result using spectrophotometry would be identical
Multi-spectral technology combines the technology of a camera with the properties of a spectral photometer. The image conveys the actual visual impression and contains the spectrum of each individual pixel. Using a dataset, both pieces of information are thus inseparable, and reliably available at the same time in different places.
A multi-spectral dataset can be visually assessed and measured at any time at a different location than where the image was made. Image capture and measurement can be divided into two steps.
Subjective and objective perception: Subjectively, the blue on the right appears brighter than the blue on the left.
The comparison is made through a "motif measurement procedure", i.e. a locally resolved measurement made without the generally conventional color control wedge. For this, an area was selected on the "Lady" motif where the analysis of the scanned data and the reference file was performed. The proof and printing of the motif were each recorded with the can:scan system.
The color quality was then assessed in the form of a locally resolved CiELab comparison.
Color coding in the images on the far right:
Black | 0 | CIEDE2000 |
Green | 1 | CIEDE2000 |
Yellow | 1,5 | CIEDE2000 |
Red | >2 | CIEDE2000 |
Data/motif Proof
Color distance to tiff reference file
Data/motif Print
Color distance to tiff reference file
As soon as multi-spectral image data are available, a locally resolved color measurement can be performed. In addition to the false-color images shown, which always depict the entire delta E color error, the three L*a*b* components can also be displayed separately in order to make errors in the brightness and the red-green and yellow-blue component visible. An example of this is depicted to the side.
For better visualisation, all errors have been scaled by a factor of 10 and the Lab* offset = 50/0/0 has been added, a medium grey thus meaning "no error".
Delta L*
Delta a*
Delta b*
Delta L*a*b*