Lost In RGB Color Space

Posted on April 23rd, 2008 | Permalink

Choosing a color profile (and by extension a color space) for your images can be a bewildering and intimidating process. I am frequently asked what profile is appropriate in a given situation, and my answers vary widely, mostly because I’m just guessing and I like variety, but sometimes because it’s an educated guess and the right answer just happens to be different than last time.

Before you tense up in anticipation of an exhaustive explanation of a dizzying array of options that… oh, you did? You can relax a little then, because for the most part, I use one of only two color profiles, Adobe RGB, also known as Adobe RGB (1998), and sRGB, also known as sRGB IEC61966-2.1 (don’t ask).

Adobe RGB has a larger color range, so I use it when the quality is important and I have time to fiddle with my images to maximize that quality. I use sRGB, which has a much more limited color range, when I need to output my images. At the end of the day, my images will usually all end up as sRGB, because most output devices (inkjet printers, computer monitors, the photo printing machines at Costco) have a very small color range, so they’re set up to use sRGB by default. For me, the question is typically not if I am going to convert my image to sRGB, but when.

I first make the Adobe RGB vs. sRGB choice when I create the image. For example, when I take a picture with my digital camera, scan something, or draw in Photoshop, I usually choose Adobe RGB, since its larger color range gives me more room in which to color correct for the best possible output. This forces me to edit each of my images in Photoshop (or something similar) and convert it to sRGB for output when I’m done, so in situations where Photoshop (and the time to use it) won’t be available before I have to output, I will choose sRGB to begin with.

Assuming the image was not created in sRGB, editing the image provides a second decision point. If the image started out as sRGB, I will almost never convert it to Adobe RGB, since the colors are already limited to the smaller color space anyway. If it started out in Adobe RGB, however, I usually wait to convert it to sRGB until I’m done editing, so that any changes I make can be done within the much larger color range of Adobe RGB. Wait, you might say, aren’t you going to lose those colors anyway once you send the image to those stupid output devices that can only handle sRGB? Yes, I’d reply, but if the colors are gonna be squashed into sRGB, starting in Adobe RGB lets me control the squashing process, instead of letting my camera or Photoshop try to do it. It requires more time and effort on my part, but usually ends up with a better product.

Before I output the image, I will almost always convert to sRGB, since that’s what the output devices use anyway. In fact, if you don’t convert to sRGB, most of the time your final product will look drab and terrible, because the output device won’t try and crunch down all the Adobe RGB colors evenly into a smaller range for you, it’ll just assume it’s getting sRGB and chop off any excessively vivid colors instead. The classic scenario of “it looks great in Photoshop but looks terrible on my blog” often occurs because the image was in Adobe RGB instead of sRGB.

There are, of course, other profiles to choose from, but each have weaknesses for which they have been eliminated from my workflow. For example, Apple RGB is widely recognized, but was designed as a one-size-fits-all for Apple color devices like monitors and printers; it’s smaller than Adobe RGB, so not as useful for editing, and larger than sRGB, so not as compatible with output devices. ColorMatch RGB is similar to Apple RGB in that it was designed to match devices (specifically, PressView monitors made by Radius), and so has the same problems as Apple RGB.

One other RGB profile that bears mentioning is ProPhoto RGB. It has a much larger color range than Adobe RGB, but until recently those extra colors weren’t particularly useful, since they were well beyond what most devices could create. These days, however, with the ability to take pictures in RAW format on many digital cameras, you can capture colors far outside the scope of sRGB, or even Adobe RGB. Even for these images, you will probably still end up converting to sRGB to output your photo, but it may be worth converting from camera RAW to ProPhoto RGB instead of to Adobe RGB, to keep all the colors intact for as long as possible. Eventually I believe ProPhoto RGB will replace Adobe RGB in most workflows, including my own, but since most devices don’t yet support ProPhoto RGB, that time has not yet come.


RSS feed of comments on this post
  1. By Dad on May 25, 2008

    Maybe this is a given and I failed to discern it as I read between the lines, the resolution on my diminutive laptop being what it is, but people do save the originals, don’t they, in case the subject that is being manipulated wins awards and will later be displayed on huge LED displays through Tokyo causing much awe and adoration to the residents and visitors of that international village? Why Tokyo you ask? The Orient has a penchant for more gaudy colors in their graphic presentations. I have yet to justify in my mind why an exact reproduction of people and places would have the colors saturated to a surreal level. The best I can make of it is they have a different reality than we do in the West.

    Be that as it may, naturally you would save a copy of the original file and squirrel it away in a safe place. No need to worry about where that place is either. Squirrels hide things and then as they dig around in places they happen to find things that they had forgotten about. Too much organization can drive a person to distraction. The important thing is to save the original and not get mired down as to where, lest you fail to do it at all. It’s kind of freeing isn’t it? And think of the joy of stumbling across that lost file years later. You don’t get that thrill if you meticulously archive and catalog your work.

    Now this all is pretty basic I realize. A person would have to be very shortsighted to fail in this area. From the experience of the first graphic artist, Thag, and his telling experience in the initial commercial venture of cave drawing we learn this lesson. It happened like this:

    Thag was in his cave having just finished cooking and consuming whatever he could scrape off of the dinosaur superhighway passing his cave. Picking up the charcoal and red clay caked stick and leaning it against the cave wall Thag witnesses the first gradient and then, well, the rest is Prehistory! The locals came to see this new expression of the almost human spirit and soon commissioned him to replicate his work in their less ostentatious dwellings. Thag, lacking the tools and experience in graphic reproduction, chose to act on his first inclination. Thag learned here that first inclinations are normally the poorest choice. He took a large rock and started chipping away at the artwork. He gathered what fragments he could for filings transfer and reproduced his composition with what he had left. The artwork was smaller and the quality acceptable. The real difficulty came when he was commissioned for a larger rendition in the communal lodge. There were just not enough fragments left to reflect the original glory that all had reveled in back in Thag’s cave. Thag’s new found wealth and fame collapsed and he was doomed to fade into obscurity as the first starving artist.

  2. By Andrew on May 25, 2008

    Hold on, hold on, you lost me at sRGB IEC61966-2.1.

  3. By Peter Pallock on May 25, 2008

    You are of course correct Dad, saving the originals (and in my case just about every permutation along the way), should be implied somewhere between those lines, which I specifically spaced out a little in CSS for that purpose.

    Any artist worth his salt (which poor Thag will tell you isn’t the best renumeration, having added thirst to hunger after the consumption of his life savings) should have an untouched digital original somewhere, just as a photographer who uses film saves his negatives.

    The option to start again from the original and rework the image for a new output device, or to fix an image that was incorrectly prepared, may be needed only rarely as said artist gains experience. However, when it is needed, having that option can mean the difference between a job’s success or failure.

    Andrew: Sorry I lost you there, but thanks for not asking. 8^)

Leave a comment

All comments from new commenters are moderated.