Singing indeed! That is an excellent outcome. Congrats. Have you tested it for colours yet?
It's been something of a struggle to get the neutrals tracking on my Epson 9800. Much less so than the Canon because of the way they mix CYM with the neutral inks. One can see this by measuring how the patch colors change when printing a target with just RGB values going from (0,0,0) to (255,255,255) w/o color management.
Much of the problem is from using DeltaE2000 v DeltaE1976. dE2k is far more sensitive to small hue and chromaticity changes near the neutrals and much less sensitive to changes in even moderately saturated colors. dE2k is a better metric to optimize because it more closely matches human vision sensitivity.
I did get better results by using the I1P target sets that Ethan's excellent post suggested, but the neutrals and near neutrals were still far enough apart that it only reduced some of the 9800's neutral variation.
So first I focused on using I1P's "optimize" process starting with the 957 default patch set to create an additional 1914 "optimized" patches. This did an excellent job improving accuracy of colored patches but had little effect on the near neutrals. After thinking about it a bit I made a set of neutral and near neutral patches that covered the entire space along the same grid spacing as the AtoB1 tables. This had patches at every position on the "neutrals" and near "neutrals" which are adjacent in the A2B1 tables. From this set I made a 957 patch, single iSis page, of these target RGB sets.
Then I just added this page to the original, iSis 1 page default set, and the color "Optimized" 2 page set to make a 3828 patch, 4 page profile target.
This resulted in a 4 page set of patches that produced not only excellent tracking in the neutrals, a real problem on the 9800, but also very good color accuracy of other, in gamut, colors.
Turns out the patch set is excellent for all glossy and semigloss type papers on the 9800.