In my quest for greater clarity about the apparent prevalence of mottly ink effects in very dark tones of prints from the 3800 (ref. Eric's image file links above) I decided to go straight to the source and discuss this issue with Epson America. They were very helpful and forthcoming, and even offered to perform a comparison test for me between the 3800 and 4800 and send me the results. I took them up on that offer and asked them to use the Atkinson printer target test image, EEM paper, their own profile and Driver, and Photoshop Determining Colors with RelCol Rendering Intent (to be comparable with stuff of my own and what Eric has been testing with). They obliged and the results appeared at my door this morning. I looked at these results under two kinds of lighting - close-up under a Solux D-50 lamp in my office, as well as in a professional light booth at a studio. Before writing this post, I called Epson and notified them of my intention to report my results here, and they had no objection.
Seen under the comparatively diffuse lighting of a professional light booth, tilting the prints in various ways, the mottling effect we have been discussing is virtually absent in both of the test prints (4800 and 3800) from Epson.
But, examined close-up under intense Solux D-50 at particular angles, the "mottling" effect appears on three prints in the blacks and very deep green in the background of the image of the rose - the 3800, 4800 and one done on an HP Z3100 (discussed below) . The z3100 print was done on Hahnemuhle Photo Rag matte, which is a relatively textured paper without brightener. Interestingly however, even under my Solux lamp I did NOT see this effect on Kayenta paper. Kayenta is not much heavier than EEM (205 vs. 192 g/m2), but it has a smoother finish than either EEM or the Hahnemuhle. Hence, I am increasingly led to hypothesize that what we are seeing is an interplay of dark ink on the uneven surfaces of both EEM and Hahnemuhle Photo-rag matte against intense light at certain angles, rather than a problem with the ink lay-down per se, otherwise one would expect the same issue on Kayenta paper, considering that I printed the Kayenta with the same K3 inks, Epson EEM profile and Epson 4800 I used to make the EMM print in my 4800 that does show the effect, which Eric also sees from his 3800 using EEM and the same inkset etc..
It has been a working hypothesis in this discussion that the apparent "mottling" is related to an excessive laydown of ink on EEM paper in the 3800 printer. Eric and I have compared notes on ink usage. This ink usage comparison is imperfect, but fwiwi, his 3800 seems to be laying down about 12% more ink per sq. ft. than does my 4800; noting however that his machine is new with much smaller sample size; as well he has used a greater variety of papers than I have. Nonetheless, one wonders even if there were as much as a 12% difference of ink laydown, whether this could cause a major difference of wetness between 3800 prints and 4800 prints. I should also mention that for the part of the Epson 3800 test in which I collaborated with Michael Reichmann back in December (as reported in his review) I do not recall the prints coming out of the machine he borrowed as being particularly wet. That is something I would have remembered, as I am accustomed to the slight dampness and curvature of prints emerging from my 4800, as well as the fact that they flatten within ten to fifteen minutes (letter size).
All this would seem to suggest several things: (1) There may be no relationship between "wet" prints and the "mottly effect" so well demonstrated in Eric's file links above. (2) This effect may be how dark ink reflects under particularly intense light at certain angles of viewing and illumination on relatively textured matte papers, due to the paper texture. (3) People getting worse results than what I am describing here (i.e. alot of wetness, apparent uneveness of ink where there should not be) may have maladjusted printers and should seek technical support.
As an aside, I was able to get printed (NOT by Epson, but elsewhere here in Toronto), the same Atkinson printer test target on an HP z3100 equipped with the latest firmware. Readers will recall there have been complaints that the Z3100 did not perform very well on matte paper and had problems reproducing reds - they were said to be orange. Well, with the latest firmware up-date in that machine, I find neither is the case, and when I compare the 3800, 4800 and z3100 target test prints, the print quality is hard to tell apart save for two things - the Epson 3800 has slightly more saturated red than either the 4800 or the z3100 (but the paper used for the Z3100 print is not brightened like EEM). The difference is mainly visible in the strawberries image, but I emphasize - the difference is slight and may be due to the paper difference. As well, the z3100 shows a more yellow yellow in the squares and stripes than observed in either of the Epsons, but in the images of the yellow flowers and foliage, all three results are fine - hard to tell one from the other. Readers unfamiliar with the Atkinson printer target test image may find this discussion more useful by downloading it here (http://homepage.mac.com/billatkinson/FileSharing2.html
) - it is third from the bottom on the right side of his downloads page.