Page 155 - Computational Colour Science Using MATLAB

142 CHARACTERIZATION OF PRINTERS
characterization is the same as in camera or monitor characterization. Device
coordinates (cyan, magenta, yellow and black) are converted into device-
independent CIE XYZ values. In this chapter the use of physical models for the
characterization of printers is described. There is particular emphasis on the
Kubelka–Munk and Neugebauer models for device characterization of half-tone
printers. Finally, in Section 9.5 two examples of printer characterization are
detailed; one for a half-tone printer and one for a continuous-tone printer.

9.2 Physical models

Characterization of input and display devices is predominantly achieved through
linear and non-linear transforms. However, although these techniques are also
often used for the characterization of printers, physical models are also
important for these devices. Physical printer models can be categorized into
two types (Green, 2002b): (i) those that aim to predict the relationship between
reflectance and dot area or colorant strength; and (ii) those that predict the
colour of different colorant combinations, in terms of either colorimetry or
spectral reflectance. It may be useful to consider these two models as processes of
colorant and colour prediction, respectively, and to recognize that they are
inversely related. Thus, many models can be used to predict reflectance or
tristimulus values from colorant information but can then be inverted to predict
colorant information.
Many printing systems print solid-colour ink in a dot pattern. Such half-tone
systems provide tonal variation by varying either the size of the dots or their
frequency. The measured reflectance of a half-tone system may be predicted by
spatially averaging the colours of the dots and the substrate on which the dots
are printed. A weighted average for each pixel in the image usually is computed
based upon the proportional areas of the dots and the substrate. Models such as
Neugebauer and Murray–Davies are used for this purpose and such models can
also take into account mechanical and optical dot gain. Mechanical dot gain is
the phenomenon where the printed dot is physically larger than it should be
because of ink spreading during the printing process. Optical dot gain is where
there is an apparent gain in the size of the dot caused by scattering the substrate.
Substrate scattering is responsible for light being absorbed by the ink dot even
when it strikes the substrate directly on an unprinted area. When more than one
colour is printed, the second colour can overprint the first. The Neugebauer
model must include the colour of the substrate, the primary colours and the
overprint colours. For a typical printing system the number of possible overprint
colours usually is quite small and therefore it may not be unreasonable to
measure them directly. In certain circumstances, however, it may be necessary to
predict the overprint colour and the Kubelka–Munk theory may be used for this
purpose (Bala, 2003).