Vampire skin tones
2010-08-18 by Allan
I recently bought a new Epson 3880 and the new Spyder3Print SR system. I'm working on a couple of iMacs, one circa 2006 and the other circa 2008. Both are calibrated using the Spyder3Elite v4, with latest s/w. I had begun using the 3880 with Hahnemuehle Fine Art Baryta paper, which is quite heavy, textured and warm. I had used Hahnemuehle's profile for the 3880, with moderate results; but I'd bought the S3P system to do the obvious, get a better custom profile for the printer and paper. I should say that I'm just using the standard Epson inks for the printer. Nothing fancy, straight out of the box. I calibrated with the EZ target and extended grays, the 4 page option. I was particularly interested in B&W images, and the results were good. My wife is quite into a form of digital scrapbooking, generating images that are a complex mixture of vector art and candid photos. When I turned to printing some of her work, skin tones were sallow and gray, pastels were truly bad. At first, I put this down to characteristics of the FAB paper. For a game, I tried a couple of prints with a Kodak Ultra Premium Glossy Paper; and since I hadn't calibrated for that, I was compelled to use a standard profile. I used an image of a family portrait that we were working on. No vector art, just parents and two children. The results were much better, with PS CS5 doing color management. For a lark, I then went to an Epson Premium Photo Paper Glossy with the Epson printer doing color management; and the skin tones were outstanding. I went ahead and calibrated for the Epson Photo Glossy Paper and did a couple of prints. The results were ghastly to say the least. Even worse than for the HN FAB paper. It appeared as if we were looking at a newly minted family of vampires from Twilight or HBO's True Blood (why are vampires so popular?) Anyway, the point is that the faces were ghastly pale, grey and artificial, especially in contrast to the excellent results from the Epson printer on Epson paper letting the printer do CM. I went ahead and added a standard warming filter to the faces in PS, just masking in the effect on the 4 faces. The print going through my S3P profile hardly showed any impact at all, even though the results were quite noticeable in PS, even with soft proofing for my profile enabled. I might add at this point that I had no fancy options set as profile options, that my profile measurements showed no bizarre errors, and that all of my print settings: media, paper handling, etc, were the same for both the S3P custom prints and the Epson CM prints. I've looked over a variety of sources on the Internet, and I can say that I'm not seeing anything outlandish about my measurement files. For the record, the measurements for the HN FAB paper targets are a bit muddier than those for the Epson Glossy Paper. As expected, white is around 97 for the Epson and 96 for the HN; likewise, black is 8 and 9 for the two, respectively. Pastel targets are a touch muddier on the HN paper than the Epson glossy, but that's to be expected, I understand. I have looked at the patches over and over again, and nothing odd stands out to me; but then, I'm new at this. Likewise, I've rebuilt the profiles on both of my iMacs for both of these papers. So, although I've used the same target prints, I've redone the work independently on two computers. Where I have noticed issues in the patches, I've gone back and remeasured in patch-reading (not strip reading) mode to get them touched up. I have also noted that my S3P sensor will often read paper white as almost black; e.g. L*=2 or so. If I do white calibration with the built-in disk, and then follow up with a measurement of the disk, L* is around 90. If I measure a typical multipurpose inkjet paper, I might get L* around 97. My papers are brighter than this, but as I say, they'll read black. If, instead, I do the white calibration not with the disk, but with the target paper's white, then I might typically get 97 for the paper and around 92 or so for the white disk. I have taken care to clean the sensor white disk with only a lens cloth; and I'm not using any measurement sets in which white calibration was done with anything except the built-in white disk. I'm just pointing out that I've had a serious degree of difficulty in achieving measurements where paper white does not read as black. This has been the case using the S3P sensor with two different iMacs that I have. As to these ghastly skin tones, I should also note that I've viewed the prints in question in both daylight as well as with a Solux lamp I have. Also, these effects (ghastly skin tones) do not show up in PS soft proofs. In fact, the soft proofs are reasonably close to the original image. It is just in making the print that the poor results show up. Having played around with this for some time now, I'm pretty sure that some of my earlier modestly good results with the HN FAB paper turned out to have been influenced by that paper's warmer white tones. With the cooler Epson photo glossy paper, the impact on skin tones is huge. Having said that, I just want to reiterate that the results with the Epson printer's native CM and the same Epson glossy paper are excellent. So if the soft proof is good, if the image is great going through the native Epson driver and printer CM, why would the prints be so far off from the proofs? If someone can tell me what I might be doing wrong here, I'd be pleased to take guidance.