Paper Testing

I got distracted with the Drobo, but I’ve finally gotten around to writing the post on paper testing that I promised last month. One of the reasons for the delay is a struggle to distinguish truth from opinion.  As I wrote, I slowly came to realize that each piece of advice falls into a continuum between two extremes: things that are so obvious that everybody probably knows them, and things that are highly controversial and have to be considered just my personal opinions. I’ve decided to order my recommendations from conventional wisdom to highly idiosyncratic (but strongly held) opinions, with more commentary as we leave conventional wisdom behind.

We’ve already covered the topic of deciding what papers to test. Let’s assume you already have a selection of papers, and you want to find out which are the ones for you. You should:

Test with the same work you’ll be printing. There is no paper that’s perfect for all images. Different works look right on different papers.

Print with the same printer you’ll be using in production. Inks and papers interact in non-obvious ways.

View the prints in the light in which they will be exhibited. Take the test prints to the most-likely exhibition space, if possible. If you’re making prints for sale, view the test prints in what you consider to be the most likely viewing environment.

View candidate prints side by side, then swap positions to catch subtle lighting differences. You may see surprising results. Light that looks even often isn’t. Use a light meter to figure out what’s going on if swapping positions causes you to like a different print better.

Test in the same size you’ll be printing. Yes, I know. You didn’t want to hear this one. It’s cheaper and more convenient to buy the little sample packets of paper and use them for testing. In fact, that’s a good way to weed out the worst papers. However, if you find that you have several papers that produce credible results, you need to do full-sized test prints, because scale affects the way a paper looks. This is especially true if the paper has any appreciable texture.

Unless the prints will be immediately framed, test for scuff resistance with a brush. You can use small prints for this. No paper is completely resistant to scuffing, but there is a huge variation. My personal bête noire is having tiny pieces of the coating flake off, leaving little white spots on the print. One of the great things about digital photography is not having to spot your prints, and I sure don’t want to go back to spotting. If you find that your favorite paper has low, but acceptable resistance to scuffing, you can still use it, but you’ll have to take special care in handling.

Watch out for optical brighteners. I wouldn’t automatically nix any paper containing them, but they will fade in time; the brighter the viewing conditions the faster they fade. They also create more variability in the viewing experience as the lighting changes, since they are excited by ultraviolet light in the illuminant.

Measure the Dmax. You can use your eyes for most things, but scoping out subtle changes in maximum density is much more easily done with a densitometer or colorimeter. You don’t need a color instrument; you just want to see how deep the blacks are. You may not choose the paper with the darkest blacks, but Dmax is valuable information to have when you are comparing papers.

Get the color right. This is hard work, and you don’t want to hear this either, but you can’t judge subtle esthetic qualities of a paper if the color is off.  By color, I mean all aspects of color: luminosity, hue, and chroma.

Let’s start with black and white images.  If you’re doing black and white work, luminosity is the most important aspect, and you probably think of it as tone mapping.  You’ve got to get that right to judge the paper properly. It’s hard to hold detail in the shadows with some papers. Other papers lose contrast in the shadow part of the tone curve, but they do it smoothly and predictably, so you can compensate in the image processing. If you haven’t worked on getting the tone curve right, you can’t tell the difference. A similar thing happens with the highlights. Some papers go from pure paper white to light grey at the first touch of ink, and some offer repeatable smooth transitions that may look crude until you’ve performed your calibration, then they look suave and delicate.

There are two main ways to make B&W inkjet prints, you can treat B&W as a special case of color, and use a color profile as your main calibration tool. If that’s your chosen approach, skip to the color section.

The other way is to use a program that’s optimized for B&W printing, or use the B&W printing tab in the driver control program. That way you’ll get a lot more black ink laid down, thus less high-chroma color ink, thus lower illuminant metamerism.  You can affect the tone curves either through the use of curves in Photoshop or directly in the controls of the printing program or the driver. There are many ways to decide how to adjust the tone curves to the paper on which you’re printing, and I can’t go into much detail in this post, but I can’t resist saying a couple of things.

The first is about measurements. Some people can look at an image of a step wedge and tell immediately what needs to be changed to make it right. I can’t; I need to measure it with a densitometer. Either way is fine, just be clear about which kind of person you are.

The second is about shortcuts. If you can find a paper that’s similar to the one you’re testing, start with the calibration settings you’ve used for it. For example, many of the baryta-surfaced fiber papers have similar tonal characteristic. You can probably get your test paper dialed in well enough for testing purposes with a few tweaks to your old formula. Later on, once you’ve decided that this is going to be one of your standard papers, you can do a full calibration.

Don’t forget that there’s a color component to a black and white print. A print with all the tones lined up smack-dab along the neutral axis doesn’t look very good. A print with the colors wandering randomly around as the tones get lighter and darker looks even worse. You need to at least get in the neighborhood of the tint you’re going to use to properly judge the test print.

Now, on to color. I’m going to assume you’re printing using a color profile. It’s not 100% necessary; given enough time and patience, you can construct a set of Photoshop curves to perform many, but not all, color corrections. I consider that method of getting the color you want to be a) old-fashioned, and b) downright masochistic, but to each his own.  Be prepared to spend a lot of time getting your test print colors dialed in.

If you’re using a profile, there are two key issues. The first is, where did you get it? The second is, how do you tweak it?

There are three common ways to get a profile.

You can download one from the paper manufacturer’s web site or use the one on the CD that came with your printer, if you are using paper from your printer manufacturer. These profiles run the (ahem) gamut from pretty good to fairly bad. They are cheap (free is the cheapest kind of cheap), and they’re worth a try. Some paper manufacturers don’t have much of a selection of profiles, so you may not find one for your chosen paper printed with your chose printer. This is doubly true if your printer is very new or very old.

You can pay someone to make a profile for you. You print out a bunch of test patches, write down the settings that you used, send the patches off to a service bureau that measures them, builds a profile, and ships it back to you. One advantage of this approach is that you get a profile that’s optimized for your particular printer, serial number 1xxxxxx. This is a real plus in printers with a lot of unit-to-unit variation, which is not as much of a problem as it used to be, especially in the more expensive printers, some of which have self-calibration features. However, the quality of the profile itself may not be as high as the printer-manufacturer-supplied profiles. The reason is that there won’t be any hand tweaking of a service bureau profile.

The most precise, but also the most expensive (unless you make a whole lot of them) and the most labor intensive, profiles are the ones you make yourself. You buy a spectrophotometer or colorimeter for between $200 and $10,000. You buy profile-making software for between included and $5,000. You print out the color patches yourself, let then age for a day or so, read them manually or automatically, depending on your pocketbook, let the profile-making software do its stuff, and use the profile it spits out. The more expensive software packages have profile editors that let you tweak your profile to your heart’s content. Works with any printer, any paper, any ink set.

The quick and dirty way to tweak a profile is with Photoshop curves. You don’t need a profile editor for this, and you can get good results if you’re at all close to start out with. The purist way is to edit the profile, and it’s more convenient to use a profile that you’ve gotten dialed in than use curves to manually make changes.

I suggest that, unless you already have the equipment and the software to make your own profiles, that you start with a profile from the paper manufacturer, if one is available. If not, try using a profile for a similar paper. If the colors are close, tweak them in with a Photoshop curve set. After you’ve decided that you love a paper, you can figure out a way to get a good profile for it.

In the above, I haven’t considered print longevity. For mere mortals, doing your own accelerated aging tests is way too difficult and time-consuming. If you find a paper/ink combination you are interested in on the Wilhelm Imaging Research web site, that’s great. If you don’t, and you love the paper, I’d say go for it, providing you’re using inks with a good permanence track record.  You have to make your peace with taking somewhat of a risk, though.

Well, I’ve talked technicalities and left-brain stuff so much that I’m starting to nod off. Let me just say that when you’ve got everything dialed in, and you’re sure you’re giving the paper a fair test, turn off the analytical part of your head, and tune in to the right side. What you’re looking for is “wow”. When your test print makes you catch your breath, you’ve found the right paper.