I second JohnnyV's procedural choice, except with some remarks:
- it is probably not helpful to select a 1% step wedge to measure the density curve. Printer dithering in general and Epson dithering in particular is good, but has its limits. As does a measurement device, and measuring relatively small steps produces too much noise. In addition, the final profile has a limited number of steps, smoothing out any device noise that may be present.
- depending on your combination of Epson driver paper selection and the real paper, the lowest L value may be reached at an RGB percentage as high as 20%. Following is a curve of GMG paper on a 4000. The green rectange indicates the lowest L value and corresponging RGB values.
- Obviously, it is not usefull to limit all RGB ink combinations to 20%, which would severely limit the maximum primary and secondary colors. However, what happens for black, happens to some extend at the primary & secondary colors as well. (Fortunately though, almost never as bad as blacks.) This means that you should try to limit the graybalance, but not the colors. This is done by selecting a different paper type in the driver.
So, to recap:
You can best try to optimize the ink deposit by selecting different papertypes in the driver. This will primarily influence the graybalance which is the significant culprit for most of the problems, while the color gamut remains relatively untouched.
Another indicator for too much ink: stray dots. You want to avoid those at all costs, because, besides that fact that it will seriously hinder effective targetprints, this can seriously glog your printer in the course of time.