If you’d like to check out how your color correction skills stack up to those of others, including some top professionals, my online discussion group recently finished a series of case studies in which group members were asked to submit their versions of some images chosen to be at least mildly challenging/interesting.

We did one of these exercises per week, for eleven weeks total. We didn’t tell the general public about it for fear of getting more entries than we could handle. But all the files are now accessible. You can read the text descriptions, but you technically have to join the group to access the images. It’s free, spamless, and cancellable whenever you like.

Here’s how it worked. An Italian member wrote me offline asking help with a photo of a brilliantly red carnival costume and mask. His problem was, it had to be printed in a book, and there is no way to achieve such vivid reds in any CMYK, let alone the one he was stuck with. And he is a photographer, not a retoucher, and so has not seen this issue many times.

I thought it would be a good exercise for the group as a whole, so I posted the RGB original and asked for people to send me their best CMYK efforts offline, together with an explanation of the procedures they followed. I then posted them and we had a good discussion of technique, some clear favorites, and requests for more such exercises. I suggested, and the group agreed, that we might as well accept that a lot of us are stuck at home with a lot of time on our hands, so that now might indeed be a good time to have such a series.

I therefore put together a series of ten original images, and each Monday I would release one. Those wishing to participate had one week to submit their entries offline, sRGB being the designated colorspace in all cases. The following Monday I would post them at the same time as I released the next one in the series.

The eleven studies drew between 21 and 30 entrants apiece. They were identified by number only without names attached (although some members identified themselves during subsequent discussions). The order in which they appeared was determined by random-number generator, since there is no better way to make one version look good than by sandwiching it between two bad ones. In addition to the online posting, I provided a downloadable .zip file containing all versions, for those wishing deeper understanding.

For each study, I quickly chose (before posting) what I thought might be the five best entrants, and averaged them into a “par” version, each of the five being weighted 20%. You may be aware from this blog that I am a big believer in this kind of blending. And if you look at these results you will be, too: everyone agreed that most of the time the par was better than any of its five parents.

Once posted, I stepped aside for a couple of days so that others could start the discussion. Within a couple of days, I posted a full analysis, discussing the strengths and shortcomings of each version and why some techniques proved successful and others not.

Here’s the announcement of the first of my exercises (reminder, anybody can read this and subsequent text messages, but to see the high-res images you need to sign up.)

The ten images I chose were selected for their diversity, and were ordered logically for the lessons to build on one another. Portraits, landscapes, flowers, night shots, scenes with action. Some were good for the workflow I advocate in Modern Photoshop Color Workflow, others less so. Some were pretty good out of the box, others had lighting issues. Eight were shot with respectable cameras, in six of which the raw capture was available. Since this was supposed to include work a retoucher might get in the real world, there was also one iPhone capture, and one desktop scan of an old, faded print. Here’s the list, starting with the one coming from the Italian photographer.

Carnival. An individual of unidentifiable gender wears a festive, floor-length gown and mask, brilliant red with yellow accents. Unfortunately, it is to be printed in a book, so it must be converted to CMYK, where it is not going to be possible to achieve such an intense red. You are given the CMYK specification; your mission is to make the costume convincingly colorful while retaining detail.

And now the set of ten selected by me, final product always sRGB.

Veiled Bride. A partially hidden face, a brick-wall background, a neutral gown and a lack of strongly colored objects come together as a recipe for boredom.

Niagara Spray. Another female portrait, the counterpart to the last one. This time, instead of being a distraction, the background is all-important, as the subject gets drenched by Niagara Falls.

Cinque Terre. A scenic hillside village in Italy, full of faded pastel buildings, with the sea in the background. Everyone agrees that the color has to be pepped up, but how far should we go?

Colosseum. A challenging night shot of Rome. How do we direct attention to the Colosseum, with so much going on elsewhere?

Panama 1978. A scan of an old print, faded. The print hasn’t been physically damaged, but the photofinishing of the time leaves a texture that needs to be addressed.

Monument Valley. Typical fodder for the workflow described in my book. A desert scene, full of subtle colors that need waking up.

A Toast to Greece. 70 years after being scattered by World War II, a family reunites in its ancestral village. The composition is fine but it’s a night shot under harsh lighting. Participants judged this one the toughest of the set.

Seated in Grass. A test to see how well we absorbed the lessons of the first two portraits. This woman is a redhead, but it’s easy to lose that color. And she has freckles, how to handle them?

Adirondacks. Autumn foliage, set off by flags of many nations, but with a twist: you’re given to understand that owing to unusual weather in the summer, this year’s display of color is the most spectacular in half a century. How to portray this without making it look artificial, or making the reds of the trees brighter than those of the flags?

Red Rose. The bookend to the first example, an image entirely about shape in the color red, this time, happily, in RGB. This is a good test of channel blending skills.