In conclusion, we learned that Adobe is indeed changing the gamut of the monitor in an attempt to match the profile of a printer. When the printer is simulated in Adobe, the gamut is shrunk. However it is not changed enough to be completely accurate to the printout which in general has an even smaller gamut. We found the HP printer profile to fit best to our printer as viewed from the delta E values measured in Lab space. The HP printer profile had the highest fidelity simulation on the monitor. Interestingly though, this fidelity did not mean that perceptually it was the best profile to use. We found that images printed with the Kodak Clear Film Printer profile looked the "best" both on the monitor and the printout even though the gamut measured on the monitor was far from what was actually printed out. It may be that because the Kodak printer uses a wider spectrum it is able to render a greater chromatic dynamic range and this is in turn more pleasing to the eye than the limited color scheme of the HP printer profile.
In order to accurately manage color we have concluded that it is necessary to use more than the generic color schemes provided by packages like Adobe Illustrator. The profiles simply aren't up to date with all the available printers. For example we believe that the HP Color LaserJet being simulated in Adobe is actually an old one and that in the years since then HP has improved the gamuts of its color printers. A start would be to make good calibrations of both the monitor and the printer and store them in ICC profiles. We learned from an editorial on the ICC web site (Tim Kohler, ICC Achievements and Challenges) that profiles have inherent problems. First of all there is no standard engine to translate the profiles. This leads to differences in interpretations of profiles and the fidelity is lost. Secondly the profiles don't deal with every scenario - things such as ambient conditions may be unaccounted for. Finally there is a lack of consistent vocabulary in these profiles. There is a dimension defined as "intents" (and this is another option in the Adobe Illustrator color settings) but it is interpreted differently by different companies. The ICC standard however seems to be a step in the right direction towards solving the difficult problem of color management.
In addition to the work presented here we explored other avenues as well. We attempted to scan images to present with our data and found that the scanner introduces even more color aberrations which the Adobe Illustrator settings cannot even begin to account for. We also looked at the spatial resolution of the various printer profiles with several test patterns and found that the color settings do not affect the image spatially.
In future work, we would look at a wider spectrum
of colors and characterize the monitor and printer over their entire gamuts.
We would also run more perceptual experiments to see what winds up looking
the best and try to figure out why this doesn't always correlate with what
is the most accurate profile.