Epson Printers Score and Fumble

Epson's Latest Printers Win/Lose Big Time

By David Em

Español

In my article Epson Inks Fade Too Fast on Some Papers in Byte.com, October 2, 2000 I wrote about the orange color shift some users are experiencing with Epson's new line of desktop inkjet printers.

The story generated considerable reader response. Most readers appreciated my bringing their attention to the problem, but a few felt I didn't go far enough in my criticism of Epson. A couple even accused me of being in collusion with Epson to cover up the seriousness of the situation.

In the conclusion to the article, I stated that if Epson was not correct in its representations about the causes of the orange shift and how to fix it, it would have a public-relations nightmare on its hands. That nightmare is now here. This month, I'll take a closer look at the problem, as well as the state of the art of printing desktop photographic-quality pictures in general. I'll also discuss Epson's new Stylus Photo 2000P pigmented-ink printer, which is supposed to make prints that last 200 years.

First, let's examine the color-shift issue. The shift occurs with the Epson Stylus Photo 870, 875DC, and 1270 inkjet printers. These printers are supposed to produce photographic-quality prints that will last between six and 25 years, depending on the paper you use. The current consensus in the printing community is that (A) the quality of these printers is indeed excellent, and (B) the ink Epson formulated for these printers sometimes interacts unpredictably with ozone or other atmospheric contaminants. When this happens, the cyan ink can rapidly fade, leaving behind mostly the yellow and magenta color inks that result in an orange image. This occurs in particular on Epson's Premium Glossy Photo paper, which it has withdrawn from the market.

It Ain't Just The Paper

The main reason I've been conditional in my criticism of Epson is that not a single one of the prints I've made with the 875 DC over a three-month period has turned orange.

This is despite the fact that I live in an ozone-rich environment and have been using the Premium Glossy Photo paper to print on. Some users have described the shift in User Groups and have posted pictures illustrating it to their websites. I don't like to evaluate situations I can't reproduce myself (this goes for sample images provided by printer companies as well, incidentally), so in the article I requested that readers who've had direct experience with the problem contact me. Several did. I asked them to mail me samples of the problem, a number of which are now in my possession.

Epson officials and Epson's customer-support line both told me unequivocally that the problem is not with the inks, and that the problem occurs only with its Premium Glossy Photo Paper. This is clearly not so. I have evidence of the problem on many kinds of paper from a variety of manufacturers. Certain papers, in particular resin-coated ones, are much more susceptible than others, but for a mass-marketed item, researching that should not be the responsibility of the end user.

Epson representatives said the cyan inks they use in this ink set are the same ones they used in the last printer line. There is a smattering of reports of similar problems with that series as well, though to a much lesser degree. However, since the new inks are rated to last about 10 times as long as the previous ones when exposed to light on display, clearly something about the inks is different. Maybe it's the formulation or maybe when the cyan merges with the other colors in the set (magenta, yellow, and black) a chemical reaction happens. It really doesn't matter. The fact is that for reasons no one yet clearly understands, the cyan inks interact with the other inks, with the paper they're sprayed on, and with ozone or other contaminants that may be present in the air in such a way that the cyan ink can rapidly fade.

Passionate Printmakers

One reason this is such a disaster for Epson is that Epson enjoys phenomenal mindshare among professional photographers and graphic designers. Where businesses think first of Hewlett-Packard for printing solutions, the creative community is by and large firmly in Epson's camp. Most of the artists, photographers, and designers I talk to are passionate about their Epson printers.

This unabashed enthusiasm percolates down to general users, who often take their cue from recommendations made by professionals. Epson's equivocations about the color-shift problem have seriously eroded the trust and confidence of the creative community. It is now up to Epson to rebuild that trust and regain that confidence.

How could this debacle have happened? Some indignant users have suggested conspiracy theories, claiming that Epson must have known about the problem all along, but that is absurd. It is absolutely not in Epson's interest to knowingly release a substandard product, especially one whose prints' long-life archival properties are a big part of its sales pitch. I have no doubt that Epson was completely blindsided by the so-called ozone problem. To get an informed view, I spoke to Henry Wilhelm, a renowned authority on photographic permanence about how inks and paper are typically tested.

Wilhelm On Testing

As it turns out, Wilhelm Imaging Research was involved in independently testing prototypes of the 870/1270 series of printers for Epson.

Some may feel that this relationship compromises Wilhelm's statements, but I don't believe that. His book The Permanence and Care of Color Photographs is the standard text in the field of photographic preservation (if you care at all about these issues it should be on your bookshelf), and he has more than once stood up to major players in the marketplace who were less than thrilled by his conclusions about their products.

According to Wilhelm, both he and the major printing companies routinely test prints for fading based on their susceptibility to light, or "lightfastness." This involves exposing the prints to a variety of accelerated light-exposure tests. New tests are being devised to factor in environmental conditions such as moisture, as well. But until the orange shift appeared, no one, including Epson, Wilhelm, or the ANSI and ISO international standards committees had ever established tests that would indicate the susceptibility of a print materials to low-level ozone or other common indoor air pollutants. According to Wilhelm, the ozone, or "gas-fading" problem was completely unanticipated.

Obviously, there's more to the fading problem than just ozone. As I mentioned above, I've not experienced the problem even though I live in Los Angeles, which has more than its fair share of ozone. On the other hand, quite a few people who live by the beach and out in the boonies have. So, a better term might be the more generic "pollution fading." This problem, unlike light and moisture, is not easily quantifiable, since local conditions vary from place to place. Additionally, electrostatic dust precipitators used in many central air conditioning and heating systems (as well as by tabletop units), or ozone emitted by laser printers and xerographic copiers can create high levels of ozone indoors. As far as testing goes, this is new territory for the entire inkjet printing industry.

The testing situation is further complicated by the fact that longevity testing is usually performed on prints that are framed under glass. The idea here is that this reproduces the typical display environment of a valued print. In this case, however, it works against permanence research because the glass prevents the paper and ink from being exposed to the suspected contaminants. Clearly, the standard testing methodology needs to be extended to include unframed prints.

Epson's advice to users is to protect all their pictures under glass or plastic, but that's a highly unrealistic scenario. These prints need to hold up under a wide variety of real-world circumstances. The fact is that most people, whether professionals or consumers, don't frame very many of the prints they make. For example, a graphic designer I know uses a Stylus Photo 1200 printer to show high-quality mock-ups to clients before they sign off on producing a project. If those pictures turn orange, the exercise is pointless. Similarly, wedding and graduation photographs are often displayed in inexpensive standup frames that don't have glass. Even if only a small percentage of prints around the world are currently being affected (what that percentage is, we'll probably never know), the risk is unacceptable.

One workaround is to print on a paper like Epson's Archival Matte paper, which doesn't seem to trigger the pollution fading problem to nearly the same degree that many glossy papers do. The problem here is that because glossy resin-coated papers give the closest physical approximation to traditional prints (which is what Epson was shooting for), that's what most people want their prints to look like. I personally love the look and feel of matte papers, but if your goal is to print glossy 8x10s, you're hosed.

Do The Right Thing

So much for the problem. What's the solution? What will it take for users to feel that Epson is doing the right thing?

Its website still advertises these inks as "Lightfast," which they are not. First, Epson must admit that there is a serious problem with these printers, and that the problem is not restricted to the Premium Glossy Photo paper. If the pollution-fading problem had surfaced back at the testing stage, I don't think this printer and ink combination would have made it to market.

One solution is to make new inks, but Epson indicates that it doesn't intend to do this. The reason, Epson says, is that reformulating the inks will require cleaning out the system heads and installing new software print drivers, both of which it says are beyond the capability of the average user to deal with. This may be true. The point is moot, however, since it is not offering that as an option. And since there is an ID chip embedded in the ink cartridges, using third-party inks is not an option either (the printers won't recognize them). So if it doesn't fix the inks, it doesn't fix the problem.

The output of these printers is so good, many users are willing to either take their chances or go the extra mile and put their pictures under glass or some other protective cover like a plastic sleeve or envelope, or even spraying a protective fixative on them. Epson should consider offering these users a major rebate on the purchase of their next model, which I suspect will solve the pollution-fading problem and be everything the 870/1270 line promised to be. Epson's pronouncements about buybacks has been somewhat inconsistent, according to some users, but when I called its Customer Help to ask about it, it offered to buy back my printer.

Epson has sold hundreds of thousands of these printers worldwide, so buybacks and rebates will undoubtedly cost it a bundle and make its stockholders mad as hell. But in the long run, it will buy back the good will of its user base, which feels as passionately toward Epson's printers as Mac users do towards their G4s.

Epson's 2000P

Since you've read this far, I'll assume that you care about printing color pictures on your desktop that will last a while.

Practically speaking, color photography has never been archival until digital photography came along. The projected display life of both Ilford's Ilfochrome (formerly called Cibachrome) and the now-obsolete Kodak Dye Transfer process is around 30 years, and not many people try making them at home. Carbo prints, which suspend pigments in a gelatin layer, can last hundreds of years, but the process is very expensive and the results are extremely fragile. The best of the current "C" prints (traditional silver-halide color prints), Fujicolor Crystal Archive, have a Wilhelm display-life rating of 60 years.

Despite everything I've said above about Epson and its dye-based inks, the current best -- and affordable -- solution on the market may be its new Stylus Photo 2000P printer, which uses pigment-based inks instead of dyes. I've been testing digital-color printers almost as long as there have been digital printers. I've worked with electrostatic plotters, dye sublimation printers, instant camera systems hooked up to film scanners, inkjets, pigmented ribbon systems, and color lasers. All these systems have achieved varying levels of success with detail, color gamut, reliability, speed, cost of ownership, noise, archivability, and messiness, but so far no single unit has solved all the problems satisfactorily.

A few months ago, I started testing the Epson Stylus Photo 875 DC, one of the printers in the series that has the controversial color issue discussed above. Then in June, I saw Epson's 2000P pigmented ink printer at PC Expo. The quality of the sample prints on exhibition looked very good, and with the projected 200 year life span of the prints, the other Byte.com editors and I decided it merited our Best of Show Award.

The 875DC

Despite my criticism of the 875DC's colorfastness, its image quality is excellent. I decided to use the 875DC as a reference point in judging the 2000P's performance. The $299 875DC runs very clean with virtually no odor, has a remarkably fast raster image processor (RIP) built into its driver, and the drying time for its prints is nearly instantaneous. I put a print in a sink under a running faucet and it didn't streak.

I was not able to get the 875 to produce a perfectly linear gray scale, meaning that it cannot reproduce certain subtle colors. The upper end of its range is very good, but the lower midtones and shadow ranges vary between a green and magenta cast. I used a natural-light balanced display stand made by Graphic Technology Incorporated (GTI) to check for color and tone consistency, and the imbalance was very apparent. Most picture information, however, is contained in the mid- and upper tonal ranges, so this irregularity doesn't show up in most photographs. When I looked at the prints in artificial light or mixed natural and artificial light, it was barely evident. In terms of color accuracy, I was able to get rich reds, but I couldn't reproduce certain blues, even though the printer uses two cyans in its six-color ink set.

I printed several test images with the 875DC including photographs of people and landscapes, and a couple computer-generated images. After a little jiggling, I arrived at a basic profile for the printer that gave me a very close match to the Eizo T960 I used as my reference monitor. After that I was able to get satisfactory first proofs very quickly. Most of the prints I produced with the 875DC are superb. The rendition of detail, saturation, and overall snap are excellent. I particularly liked the results I got with Epson's Archival Matte paper. Images printed on gloss and semigloss papers were similarly bright and true. To my eye, these pictures are equal or superior to any chemistry-based system I've worked with in the past.

Here Today, Here Tomorrow

Like the 875DC, the pigment-based $899 2000P is a six-color inkjet printer (2 cyans, 2 magentas, yellow and black). There are pros and cons to printing with pigmented inks. The big pro is that certain pigments can last hundreds of years, whereas dyes tend to fade on exposure to light much more quickly. Jan Van Eyck's paintings, which were painted with pigmented oil paints on boards and varnished on both sides half a millennium ago, are as bright today as the day they were painted.

The 2000P's inks are encapsulated in resin, making them impervious to airborne contaminants. Wilhelm rates their life expectancy on display before noticeable fading occurs at 140 to 200 years, depending on the paper they're printed on. On the down side, pigments tend to have a narrower color gamut, and are much more difficult to disperse in a liquid medium like the tiny droplets used by inkjet printers. It's also much harder to achieve a uniform high gloss surface with pigmented inks.

When I ran my gray scale tests on the 2000P and looked at them under the GTI display's daylight bulbs, I immediately saw that there was a distinctive green cast to the image. The phenomenon of colors appearing differently under different light sources is known as Metamerism, and I've rarely seen a clearer example of it. Darker ranges of the black and white spectrum also skewed towards a reddish tone. I printed several pictures with a lot of color in which the green cast had little observable effect. But in images containing large areas with neutral tones, and in certain shadowed areas, I wasn't able to compensate for it. In normal room lighting the problem is minimized, but in broad daylight it's an issue.

I had to work much harder to establish a basic setting for the printer than I did with the 875DC. I can look at a picture on the T960's screen and know exactly what to do to make a great print with the 875, but getting a good print of the same image from the 2000P takes considerably more finessing.

As I expected, the color gamut of the 2000P is not as wide as the 875DC's. I had a lot of trouble nailing certain yellows accurately, and light rose tones consistently came out violet. I also wasn't able to get the reds to come out as saturated as I would have liked. The 2000P uses the same inks as Epson's large format 7500 and 9500 printers, from which I've seen spectacular results, so I'm not sure that I've pushed this printer to its limits yet. I probably need to isolate and tweak various picture elements by hand on the software side with Photoshop and see what happens. I have managed to pull some gorgeous prints with this printer, however.

Black And White

If you're a lover of black and white photography, as I am, your best bet for now is to set up a separate printer for your monochrome work.

Printers like the 875DC and the 2000P convert monochrome information into CMYK (Cyan, Magenta, Yellow, Black) color space, so you never get a true black and white gradient, you get an approximation. Usually the approximation is less than ideal. (The 875 has a black-and-white setting that's supposed to yield noncomposite black, but the black ink must have a natural magenta hue to it, because I wasn't able to make a completely neutral print.)

The solution to the problem is to use a monochrome ink set. Lyson sells a very nice set of warm, cool, sepia, and neutral inks as well as a new "Small Gamut" hue-controllable monochrome inkset. Another candidate is ConeTech's Piezography quad black system. Both Lyson and ConeTech have engineered their inks to work with certain Epson printers (though Epson may invalidate your warranty if you use them). The samples I've seen from both systems look promising, and they both offer long life. I'll be running some tests with both of them after the start of year.

The Future

Very few photographers have set up color darkrooms on account of the expense and complexity of the endeavor. The digital darkroom changes that situation forever. The image quality of prints made on inexpensive digital desktop printers is already superior to traditional silver halide prints, in my opinion. It's likely that the missing link to no-compromise, long lasting digital prints will ultimately be provided by pigmented inks and improved papers that are specifically formulated to work with each other.

During the next month or so, I'll report on new systems being introduced at Comdex/Fall 2000 by Epson, Hewlett-Packard, and others. And I predict that a year or so from now, professional photographers, designers, and artists, as well as just plain folks, will be routinely printing inexpensive color and black-and-white photographs that will outlast us by a good lifetime.

This article originally appeared in Byte.com.
You can read more of David Em's articles on multimedia tools at http://www.byte.com/index/MediaLab.

You can contact David Em at: davidem@earthlink.net