Mark, I am not sure if I know how to analyze a print well enough to know that I would need a custom profile. How would I know it 's not my printer? or my calibration ? or ?
Go to my most recent article on the Epson 4900 (the Review Addendum
http://www.luminous-landscape.com/reviews/printers/epson_4900_printer_review___addendum_february_2011.shtml ), click on the link to Outback Photo under Figure One and download that target which is the same as my Figure 1. Save it to your hard drive.
Calibrate and profile your monitor. If you are working in a dim environment and likekly to view the prints under D50 illumination, a reasonable starting point for your calibration points would be White Point of 5000, luminance of 110 cd/m2 and gamma of 2.2 or L*, depending on the options in your calibration package. These are suggested start values, which may need to be adjusted to your environment.
Now, with all that set-up, load your Moab profile into the Photoshop or Lightroom Print dialogue, open the printer test image you downloaded, make no adjustments to it, make sure there is no colour management active in the printer driver, check that the correct Epson reference paper corresponding with the instructions for using the Moab profile is selected in the printer driver and print it on an 8.5 by 11 inch sheet using Relative Colorimetric with Black Point Compensation, or Perceptual depending again on the advice coming with the profile. Then turn on Soft-proofing, making sure the soft proof set-up corresponds with the profile and other printing conditions you are using. Compare the print (well lit) with the softproof and see how good the matching is to the display. You will be able to see immediately if there are major problems with luminosity (brightness, contrast) or colours and where these problems are most pronounced. Depending on the nature of the disconnects, it is then possible to do more diagnostics to figure out whether the problem is the display or the profile, or a mixture of both, but get to first base on the initial set of results, and see what happens. If it's all good, we needn't get into diagnostics.