I’ve written articles in the past about color management—you know, ICC profiles, calibration, metamerism, blah blah blah—but the world keeps changing around us. In general, things change for the better, which is great, but as a photographer looking to display images on the web, there are always things to watch out for.

I found a great article on the burgeoning “Gear Oracle” site about web browser color management, which breaks down the effects of displaying images with different color spaces in different browsers.

It’s not all great news, but it’s important to know how things work so that you can mitigate any potential problems. Color management is basically a briar patch no matter how attentive and diligent you are, and trust me, I know this from close personal experience printing photographic reproductions for five years.

Read Web browser color management guide via Gear Oracle

Have you ever wondered about that “gamma” thing you keep seeing? You nerd.

Really, though, gamma is important and you have probably seen the word all over the place in photography and design. It’s actually a really cool thing and when you understand how it works you will likely feel better about yourself, your photographs, and about the universe. Well, you’ll feel smarter, anyway, and you will be. You’ll also be able to add another item to your lists of:

1. Answers to questions nobody will ever ask you,
2. Greek letters you recognize, and
3. Awkward things to bring up on a first date

You can already check off number two if you look up on the right there. Yup, that’s gamma.

Additionally, if you are friends with other photographers and they don’t know what gamma is or how it works, you might come out of this looking like a rockstar. At least to the extent that rockstars are knowledgeable about non-linear power-law expressions.

So what is gamma? Aside from being the third letter in the Greek alphabet and a type of brain wave, the word “gamma” is used in imaging (photography, design, broadcast technology) to refer to “gamma correction,” which relates to adjusting the luminosity of an image as it is displayed on (usually) a screen of some kind.

You have probably heard or seen things like “gamma 1.8” or “gamma 2.2” thrown around, especially in the photography world, and that’s where the math starts to creep in. But before I get all numeric on you, let’s take a look at what gamma correction really means, how it’s used, and why it’s important to you.

Gamma Correction, What Does It Mean?!

The truth is that gamma correction is pointless. Gamma correction is one of those things that was invented to solve a problem (and just in time, too), and then its use spread throughout the world and was written into standards that were carried through generations of technology until we reached the point where it wasn’t necessary… But it was already too late. So even though gamma is a vestige of a problem we more or less no longer have, we can’t stop using it or we’ll create even more problems.

Gamma correction is a way of adjusting the luminosity of an image in a non-linear way, which means that the change in luminance for a particular value in a source image depends upon that value. Uh oh…

Here comes a graph, look out!

Gamma 2.2

Now, the first question you should be asking is Why would you want to change the luminance values in an image? and the answer is quite simple. CRT technology.

Remember CRTs? Cathode ray tubes? Those big, heavy, glass picture tubes used in every single television set and computer monitor from around 1922 when they were first commercialized until about 2007 when LCDs first overtook CRTs in overall sales?

This is what a cathode ray tube looks like when it's drawn in MS Paint by a crippled giraffe.

They had a pretty good run, I’d say. Well, cathode ray tubes have what they call a “triode characteristic,” which basically means that the relationship between the input voltage (incoming image luminance) and the luminance on the screen itself is not linear. In fact, at low voltages, the luminance is far too low, and as the voltage increases the luminance does not quite increase in step.

This is demonstrated by the green line in the graph above. The “input voltage,” or “source luminance,” is on the x axis, across the bottom, and the luminance seen on the screen, or “output luminance,” is on the y axis, up the left side.

You can see that the output luminance is severely depressed, and you would be, too, if you were trapped inside a cathode ray tube. Anyway, to make a long story only slightly shorter, this is a big problem. This is a problem because you can’t just brighten the entire image or all of your very light areas basically get overexposed and wash out. What you need to do is “reverse” the effect of the triode characteristic… Which is exactly what gamma correction does.

The red line is the mathematical inverse of the green line. If you process your input image according to the red line and display it on a CRT, the image’s luminance will even out to the black line, which is what you want. A perfectly linear relationship between the image data you send in and what you see on the screen. Brilliant.

The Math

Let’s just talk about math for about one minute more. Remember equations? Right. An equation is what mathematically describes your output luminance based on a particular value of your input luminance. For gamma, it’s pretty simple:

Output = Input^gamma

The output luminance is equal to the input luminance raised to the power of the gamma value. In actual textbook mathematics of course they would use the actual Greek symbol “gamma” up there, but this is a photography blog. To get the opposite curve, you just replace “gamma” with its reciprocal, “one over gamma.” So, for the curves up above, this is the equation I actually used:

Output = Input^2.2

The infamous 2.2 gamma! Indeed. So when folks talk about a “gamma of 2.2” or a “gamma of 1.8,” what they are talking about is how much luminance correction is being done. You’re learning so much already! And I’m not even done yet!

What Gamma Setting Should You Use?

Powerrrr!

Once upon a time, when early Macs were the first home computers to have color screens and the ability to view and print color images, Apple engineers encountered this triode characteristic. Expectations were a lot lower then when it came to computer graphics, but when someone opened a picture on their snazzy Mac and then printed it on their equally snazzy color LaserWriter, they expected it to come out looking more or less the same as it did on the screen. But this was hard to pull off when the screen was lying to them about the luminance values.

Back then, Apple was not concerned with how accurate the color was in a universal sense—in the ICC color profiling sense that would come much later. They just wanted the images to look the same when printed on the color LaserWriter. So, they did some experiments and they found that a gamma correction of 1.8 would get it right in line with what came out of the printer.

A bit later on, another group of bright folks got together with a different purpose. They called themselves the National Television System Committee, or NTSC for short. The FCC created the NTSC in 1940 to standardize black and white broadcast television. By 1950 they were getting together to standardize color television. Ultimately they developed what is still referred to as the NTSC standard (or NTSC color television standard, sometimes), and saved the world of television. Alright, not really, but they certainly knew all there was to know about TV and they were intimately aware of the gamma problem.

Their solution was to use a gamma correction of 2.2 to make TV images look correct on your home screen in your darkened living room. When Microsoft Windows came out and started pushing more color computers into the home, they adopted the NTSC’s 2.2 gamma recommendation.

With the differing gamma values of 1.8 and 2.2 being used by Macintosh and Windows-based computers, an image that looked correct on one would often look too bright or too dark on the other. This is the way it was all the way up until last year when OS X 10.6 “Snow Leopard” was released and for the first time ever set the default gamma on a Macintosh computer to 2.2.

So What?

You just read 1,300 words about gamma correction, viewed a pretty graph, and learned a bit of history. So far, you have a handful of Trivial Pursuit ammunition and a radical ice-breaker for your next blind date, but none of this really helps you in your quest to be the world’s most awesome photographer (although reading my blog is always a good first step!)

The most common question asked about gamma is, “What gamma setting should I use?” Since OS X now defaults to 2.2 and Windows has used 2.2 for decades, the answer is 2.2. You will probably not gain anything by using a different value unless the lighting conditions where you’re sitting are totally extreme. Like the beach in July. Or a cave deep beneath a granite mountain (which is where my evil lair is located).

Thanks to color correction systems, ICC profiles, and soft proofing, you really don’t have to worry about gamma too much. Images in Photoshop (and Firefox 3.5 as well) will be adjusted based on their embedded profiles (assuming they have profiles).

The Most Important Thing

The one thing you need to remember is this: when you are exporting photos for use on the web, always, always, convert them to the sRGB colorspace and tag them with the sRGB profile. The profile itself contains a gamma setting (typically, guess what, 2.2) and folks on the web using Firefox will see the image properly adjusted while folks with browsers that don’t support color management will see the closest to what you would expect. sRGB is very reliable when it comes to this.

If you have any questions about gamma, if I didn’t explain something very well, or if you want to lavish me with praise, please leave a comment. I do read them. I really do.

Even with the best equipment that money can buy, ICC profiles, spectrophotometers, an iron-clad color management workflow, and a high-end monitor, your eyes are the ultimate judges of your work. But eyes, they don’t work alone; you can’t see anything without light, and the quality of the light will have as much an effect on what you see as the color of the print itself.

I got onto this topic after reading Michael Johnston’s overview of his Viewing Station. All these years I’ve been experimenting with lights in my studio space, let’s call it Single-Serving Photo HQ—or, as my friends call it, my bedroom—and I never once thought to write about it.

After the jump I’ll tell you how to dramatically increase your viewing conditions for about $15.

Michael says he uses one of those clip-on Verilux full-spectrum fluorescent setups. When I started down this path, I hadn’t heard of Verilux specifically, but I was sure I wanted to go fluorescent after being turned off by these so-called “daylight” incandescent bulbs that were nothing more than regular bulbs with bluish glass designed to offset the orange light they actually generate. Michael’s entire lamp, which includes the bulb, cost him around $80. We can do better than that.

Get a Bulb

I wound up sampling several compact fluorescent bulbs from 1000bulbs.com and eventually decided upon a 100 watt equivalent 5100k model. Here’s a picture of what it looks like, sort of:

There are perhaps four important considerations when looking at bulbs for your viewing environment.

1. Wattage (or equivalent wattage, when talking about fluorescent)—I like my viewing conditions to be fairly bright, so I went with 100 watt equivalent, but you may prefer to go higher or lower.
2. Spectrum—Get a bulb that is classified as “full-spectrum,” meaning that it doesn’t purposefully exclude certain wavelengths.
3. Color temperature—Anything from 5,000K and up should do fine. I prefer 5,100K, you might go as high as 5,400K.
4. Color Rendering Index (CRI)—This method of measuring the color accuracy of a light source has its flaws, but it’s better than guessing; a higher number is better, 100 is perfect.

A Note About CRI

The Color Rendering Index (CRI) is a measure of a light source’s ability to reproduce color. Those of you mathematically inclined or simply curious can read about it on Wikipedia. This particular system has its detractors, but it’s the system most widely used at the moment, and 1000bulbs.com lists the CRI value for most of its bulbs, especially the ones billed as “full-spectrum.”

The Verilux bulbs that you find in systems such as Michael’s rate around 86 and up on the CRI scale. The bulb I chose scores an 82, which by all accounts should be good enough for any normal person. Some halogen bulbs are rated 100, which is as accurate as a light source can be on that scale, but halogen is expensive, hot, and may be too bright for some people.

23 Watt, full-spectrum, 5100K CFL from 1000bulbs.com: $5.71

Get a Lamp

What good is a bulb without a lamp to screw it into? Perhaps you have a spare lamp somewhere in your house that you can use, but if you don’t, do not despair. I went down to the home improvement store and picked up a simple work light (sometimes also called a painter’s light or a clip light). Here is a pretty basic one offered at Lowe’s; I don’t remember where I got mine.

Simple clip-on work light: $7.48

Clip it onto something, such as your apartment’s fashionably painted structural beams, and away you go!

I actually bought three bulbs, three work lights, and plugged them all into a single power strip so I can turn them on and off with one switch and light my whole computer area. It’s very helpful to be able to lay prints out on top of the printer or on my desk and have the same quality of light everywhere.

You will definitely see a difference in the appearance of your prints as you move them from one light source to another. Nearly every print will have some variation in the way it reacts to light. Some inkjet inks are known to be rather more prone to this change in appearance (which is called metamerism), such as Epson K2 (the inks I use). It’s not necessarily a bad thing, but as an artist you need to be aware of the possible ways in which your work will be viewed and whether you are satisfied with the work’s performance.

So there you go. A full-spectrum viewing solution for less than $15.

Though my question has little to do with your most recent article, I find that the expertise you’ve shared since I began following your site to be compelling and hope that you can provide me an answer that will serve my needs. I have seen many digital photos over the years, some pretty dismal and some pretty spectacular. My question was born out of seeing, for the first time, an exhibition comprised of 150 works by Ansel Adams, which were nearly all “Silver Gelatin” prints. My question is (setting aside for the moment the composition, line, form and majestic beauty of many of the locations) can any digital print be made in such a way that a knowledgeable observer would not be able to distinguish it from a print made using the silver gelatin method?

In the interest of full disclosure, I should probably mention that the comment was posted by my father, who, having a degree in fine art and a gallery of his own, takes a vested interest in some of the topics I discuss. That said, I thought it was a very relevant question and one that many of you might have thought about, too, so here’s what I think.

First, it would depend on how knowledgeable the viewer was and how close they could get to their subject. Second, there are many technologies available, so the short answer is “probably,” but the long answer is more interesting.

So-called “lightjet,” which combines a digital (laser-based) exposure of photographic paper with traditional (chemical) development methods, produces very fine results, especially of full-color images. Upon very close inspection, however, it would be clear that there are many colorful dots making up the print. Still, these prints have fantastic longevity, are often quite lustrous, and are available for a lot less money, comparatively, than other methods. An added benefit is the ability to print on any brand and type of photographic paper available to photographers, from Kodak Endura to Fuji Crystal Archive. There is a certain je ne sais quoi surrounding real photographic papers that may be the strongest argument for lightjet.

Inkjet prints boast a longevity nearly comparable to traditional development and are capable of a much broader range of color than lightjet. The highest-end inkjet printers now deliver between seven and twelve physical inks in picoliter droplets that mix on the paper to create a continuous tone image. The ink droplets are dispensed by a piezoelectric system and can be either dye-based or pigment-based, each having their own archival and color properties. Inkjet printers, however, are much more expensive to run than ordering your prints from a third-party photofinisher (who probably uses lightjet), and if you need the highest quality available, you will have to buy, configure, and operate the setup yourself, which is no small task!

Giclee (zhee-clay or gee-clay) has also held its own against the influx of lower-cost inkjet solutions, boasting the ability to print on materials such as canvas and at resolutions beyond what inkjet or lightjet typically can achieve. Giclee (sometimes called Iris printing because one of the original models was called Iris) is essentially a CMYK inkjet system, meaning that only cyan, magenta, yellow, and black inks are used, though I have heard of giclee printers that use six inks. The ink is fired from glass nozzles at one million droplets per second and each drop is electrically charged so it can be directed toward or away from the paper by electromagnetism. The paper itself is affixed to a drum that spins at about 180 inches per second. Giclee is probably one of the most mechanically impressive printing methods around.

Artists have chosen giclee for years because of its faithful color reproduction and ability to print on “artistic” substrates such as canvas. A single giclee print, however, can cost $50, $100, or $200 to produce, not including the calibration and other services required to achieve the results you need. Giclee is more favored by painters than photographers.

Four-color offset lithography, which is how all print publications are produced nowadays, is actually capable of near-giclee quality, however American print shops tend to be too traditionalist to adopt the color management methods necessary to produce fine art prints to an exacting standard. Bill Atkinson, a man absolutely fanatical about color accuracy, collaborated with a Japanese print shop to implement color management methodologies for their four-color presses. In return, they printed his book of rock photographs. That book may be the only example of accurate color reproduction through offset lithography on any American bookshelf. The difference between the capabilities of high-end offset lithography and giclee is entirely due to the willingness of the technical staff involved to use modern digital color management methods.

At the end of the day, can any of these digital solutions deliver a result as austere and striking as a pure black and white gelatin silver print? Probably not, but they can come very close. I am convinced that Ansel Adams himself would be a dedicated and outspoken advocate of digital photography and all of its methods were he still alive today.