The Applied Color Theory list, which discusses matters of interest to my students, friends, and colleagues, has been operating since 1999. For almost the past twenty years it has resided at yahoogroups. After many years of deterioration, yahoogroups recently decided it would now become e-mail only, no images allowed.

We have therefore moved the group lock, stock and barrel to a new home at groups.io. You are welcome to join the over 2,000 members who in discussing or reading about color. The transition took place earlier this month and the list has become quite active, with threads on luminosity masking, blue filters in eyeglasses, my forthcoming book, and others.

We haven’t implemented everything yet, but among the important changes is that the group is now public access—you can read posts without becoming a list member, though you can’t add to them.

{ 0 comments }

Previous postings evaluating how the Picture Postcard Workflow did compared to the retouching team on the MIT dataset established the obvious, that neither PPW nor any other workflow is the perfect solution every time. The question is, how do identify when PPW shouldn’t be expected to be better?

Sometimes the image simply doesn’t have enough potential for improvement; sometimes it can be improved but the tools of the PPW offer no particular advantage; this typically is when an image has lots of natural color variation and no need to emphasize detailing in the brightest objects. I’ve shown examples of all three in previous posts in this series.

But these are predictable: I look at them, and expect nothing special out of PPW. The reverse is an image of a canyon or desert. These play into the hands of PPW’s ability to differentiate similar colors. I expect to do much better than the MIT retouchers when confronted with this type of image, and this was indeed the case, as shown here.

This post is about a learning experience: the type of image where I expected to win as easily as with a canyon shot, but it turned out otherwise. Sometimes that was due to sheer stupidity, which is an occupational hazard and examples of which are in other posts.. But in a few cases I was confident I would win and did not, even though I made no obvious mistakes. That category is what I’m about to discuss.

The first surprise had to do with size. PPW has a big advantage in portraits and other fleshtone images because it can create color variation without changing the overall perception. It can, for example, make lips and cheeks redder, compensated for by making certain areas cooler. This is a decisive advantage when the face dominates the image. I would have thought the advantage would persist at least somewhat in smaller sizes. But when the face was small because the subject was portrayed full-figure, if PPW had an advantage it was for other reasons. (The set contained no nudes, which probably would have been wins for PPW.)

The second involves a specific color. I expected no problem winning the competition on the following two originals.

4897-default and 5054-default: I expected better results than I got from PPW on these two originals.

4987-default and 2764-default: I expected better results than I got from PPW on these two originals.

Both are dominated by various shades of a single color. PPW usually has a big advantage in such images because of its ability to drive colors apart. And indeed, PPW won easily in the other six portraits where the faces, as here, were large. But not here, and not in the picture of the church choir stalls, either. Maybe they could be called a tie, but certainly no clear superiority.

It is not a coincidence, in my opinion, that this was the only similarly-sized portrayal of dark-skinned African-Americans, For some reason, the color variation that PPW invoked was not as attractive as in other cases. Here’s the comparison.

4987-PPW: Plenty of snap, but perhaps too much hue contrast.

4987-PPW: Plenty of snap, but perhaps too much hue contrast.

4987-par: The average of the retoucher versions. Too flat, but if we consider color only, this is possibly the better of the two in spite of being too yellow.

4987-par: The average of the retoucher versions. Too flat, but if we consider color only, this is possibly the better of the two in spite of being too yellow.

I don’t think the par version is very good. It’s flat compared to the PPW, and too yellow. Nevertheless, if I had to choose on the basis of color only, I would take it. I don’t see an obvious mistake in the PPW color but somehow it’s not attractive. Based on a similar result in the church image, I have to suggest that when the base color is more neutral we have less tolerance for color variation than when the color is more saturated.

Skin is basically red in every ethnicity, but light Caucasian skin is so saturated that a PPW routine can never yield anything that no longer is red. Certain parts of the face may get cooler, but that just makes them a duller pink, which we are inclined to accept. But when the skin is a deep brown these same parts of the face may become neutral or even green or blue.

The third surprise was how seemingly irrelevant objects can affect the need for variation. Here’s another pair of originals:

0017-default and 2796-default: two images that feature green. Will they handle the same way?

0017-default and 2796-default: two images that feature green. Will they handle the same way?

I again anticipated an easy win for PPW on both images because PPW thrives on large areas of green. That’s how it turned out on the first:

0017-PPW: the greens have plenty of detail and also hue variation.

0017-PPW: the greens have plenty of detail and also hue variation.

0017-par: the average of the retouching team is not competitive.

0017-par: the average of the retouching team is not competitive.

I assumed that the other original would handle just like the above, but obviously I was mistaken. Here’s my version compared with the best result from the five retouchers:

2796-ppw: plenty of hue variation between the soil and the greenery, but who needs it?

2796-ppw: plenty of hue variation between the soil and the greenery, but who needs it?

2796-C: rich, sunny colors make for an attractive presentation.

2796-C: rich, sunny colors make for an attractive presentation.

My takeaway from this is the viewer is subconsciously attempting to differentiate the browns from the much larger areas of green, rather than finding more differentiation in the greens themselves. In the other example, although some browns exist, they are apparently not the priority. Another possibility: in the second image the greenery could be described as boring. In the first, it has interesting features that we are apt to want to investigate. And maybe more important: if the only question is how different are the browns from the greens, then PPW wins. But that’s not the only question, it’s not even an important question, because there’s plenty of differentiation in the original, no need to exaggerate it.

These theories may or may not be correct, but still, I learned a lot from these unpleasant surprises and hope that you may as well.

{ 3 comments }

The college basketball playoffs in the United States have just ended. Basketball sets the theme for the apparent resolution of a longstanding issue involving photography, and by implication many other areas of design.

If without permission, you re-use someone else’s photograph, you may be looking at paying for copyright infringement. This is assuming that you somehow got hold of a digital copy of the photo.

But what if you do not literally copy? What if you look at the photograph and decide to set up a shoot that will duplicate it closely? Is your photo, clearly intended to look like the other, an infringement? An urban legend states that it is. The three leading cases, however, say that it is not. This differentiates photography from music, where copyright infringement has been found where one piece sounded somewhat like one written by a different composer, even where there is no proof that the “copier” knew of the existence of the first piece of music.

In the three cases involving a “copied” photograph, including the one I’m about to discuss, the copier most certainly did know, and did intend to steal the concept. In the most well-known previous case, Skyy Vodka, whose signature blue bottle is quite attractive, was in talks with a photographer about an advertising campaign. He provided them with three sample shots that they liked very much, but they could not agree on a price to use them and in any event the photographer insisted on a licensing arrangement rather than a total transfer of rights. The company then found another, cheaper photographer, to whom they showed the first photographer’s work, saying, this is (wink, leer) the sort of thing we would like. The cheaper photographer obediently produced a series of very similar shots, which Skyy used in advertising. The first photographer then sued for copyright infringement.

He lost. And it took him ten years. The first judge, ignoring all the skillful lighting and background effects, ruled that the photographs were uncopyrightable because too much of them depended on the vodka bottle, which of course the photographer had no claim to. Having determined that no copyright existed, he had no need to rule on whether the second photographer’s work infringed it.

An appeals court, however, decided that the photos could be copyrighted and told the first judge to consider the infringement question. He did, and concluded there was none. He seemed to agree that one independently-shot photograph could theoretically infringe another’s copyright, but he said that in the vodka bottle situation “highly similar” was not enough, even granted that the second photographer was trying to copy the first one’s work. The two photos needed to be “virtually identical” for the plaintiff photographer to win, and of course they were not.

A second appeals court agreed. Here is the opening of their opinion: “This long-running litigation is fundamentally about how many ways one can create an advertising photograph, called a ‘product shot,’ of a blue vodka bottle. We conclude there are not very many.” This was in 2003, ten years after the original photographs were taken.

In other words, the rationale for requiring that the two be “virtually identical” is that otherwise too many legitimate attempts to portray the bottle could be challenged. That ruling left open the question: what if there is no doubt that the concept of a photograph was copied, but the photograph was so creatively arranged that nobody would ever come close to it by accident?

Just such a photo was taken in 1984. The photographer’s contribution was to instruct a college student named Michael Jordan, wearing his USA Olympic sweats, to assume a ballet-like pose while soaring to dunk toward a basket that the photographer had constructed to seem impossibly high. The whole scene is outdoors, not on a basketball court. It appeared in Life magazine and aroused the interest of Nike, which paid the photographer $150 to use it in a slide presentation.

The Jumpman logo.

The Jumpman logo.

Shortly thereafter, Nike decided that it needed a picture like it for its own marketing purposes, but Jordan had to be wearing Nike shoes, and also the attire of his new team, the Chicago Bulls. So they hired Jordan to duplicate his pose, leaping outdoors toward a basket that was at a more reasonable height than in the original, and hired a photographer to shoot at the same unusual angle. This new photo was used over and over to promote Nike’s fabulously successful Air Jordan line of shoes. Furthermore, they derived a bitmap image, a solid silhouette, from the Nike photo. This was used as the logo for the fabulously successful Jumpman line starting in 1987.

When the photographer saw the new Nike picture, he threatened to sue, so Nike paid him $15,000 for the rights to use its own image for two years, although apparently they continued to use it for a while afterward.

Sales of the Jordan-branded lines were in the billions of dollars. The original photographer thought it would be useful if he shared in some of the largess, so in 2015, more than 30 years after the photo was taken, he sued Nike for copyright infringement. The federal judge hearing the lawsuit threw it out without a trial, stating that neither the second photograph nor the logo was an infringement as a matter of law. In doing this, he relied not only on the vodka-bottle case, also on an earlier one involving a student of mine, who specializes in photographing babies. She had done a beautiful magazine spread in the 1980s, featuring nude babies apparently floating in air, carefully lit and tastefully presented.

A midwestern hospital thought that this would be just the thing for one of its brochures promoting its neonatal care facilities. It therefore contacted my student about purchasing the rights. Before anything could be finalized, however, objections were raised that her photographs showed the babies’ genitalia, which would no doubt be shocking to the hospital’s audience. They therefore hired another photographer to take extremely similar photographs, more discreetly posed, of different babies.

My student sued, and lost. The court acknowledged that the hospital’s intent was to copy her work, but informed her that designs, concepts, and lighting methods are not copyrightable, only her specific implementation with specific babies. So, the result again: the allegedly infringing photo needs to be virtually identical to the original or the first photographer is out of luck.

 

The 1984 original photograph by Jacobus Rentmeester.

The 1984 original photograph by Jacobus Rentmeester.

The case that was just decided is more complicated because, among other things, the same individual was photographed each time. Still, there is no question that the following things cannot be copyrighted in and of themselves.

*The particular pose, unusual as it is.

*Michael Jordan dunking a basketball.

*Michael Jordan playing basketball outdoors.

*The angle the photographer chose for the shot.

*A basketball hoop much greater than the regulation height.

However, if enough of these elements are included simultaneously, then copying all of them could, according to the court, result in a copyright violation. The Nike photo copied four of them. The logo didn’t copy any of them: it’s an unidentifiable silhouetted figure, based only on the Nike photograph, where the pose is slightly different than the original. Therefore, all the judges hearing this case agreed that the photographer had no valid claim against the logo.

The Nike photo, admittedly inspired by Rentmeester's.

The Nike photo, admittedly inspired by Rentmeester’s.

But the Nike photograph itself? Well, in addition to the lower basket, the background is totally different, as is the clothing. Jordan is older and the pose is, understandably since it is taken mid-air, not quite identical to the original. There’s a colorful sunset on the left, whereas in the original the sun is not colorful and it’s on the right. Jordan is much larger compared to the background in the Nike shot. The original focused the lighting on the ball, while the Nike version focused on Jordan’s face.

For two of the three appeals court judges, that’s enough difference. Their holding: if a nonexpert viewer can readily tell the two photos apart, then there’s no copyright infringement. Nike wins.

The third judge agreed with the others that the original photographer had no claim to the logo, which was where the big money was. As to Nike’s photograph, though, he would have sent the case to a jury to decide whether the two images were “substantially similar”.

Since the ruling wasn’t unanimous, I thought it might attract the attention of the Supreme Court. But no, they did not wish to take the case, which means it is now over, ending as the other two did with a loss for the photographer.

The lesson? Suing someone on the basis that a photograph, though admittedly not a direct copy of yours, is too similar, is not going to succeed even if it’s clear that the other photographer was intending to pirate whatever creativity you put into it and hope that his work would be mistaken for yours. All these photographers lost a lot of time and money chasing rainbows.

On the other hand, it is expensive to get sued, even if you eventually win. These three copiers at least tried to make a deal with the original photographers. That made their defense a lot simpler, and would be a wise choice for any photographer or designer in the same position.

{ 0 comments }

R.W.G. Hunt 1923-2018

by Dan Margulis February 26, 2019

I recently learned of the loss of an outstanding contributor to the practical side of color science. Robert W.G. Hunt died in October at age 95. He was a prolific writer, but he is best known for his massive text The Reproduction of Colour, now in its sixth edition. Most of it is quite geeky, […]

Read the full article →

Image Distribution Surprises (The MIT 5k dataset 8)

by Dan Margulis December 7, 2018

The MIT study offers a unique opportunity to study the distribution of images. Most of the images that I and other authors collect for our own purposes test or prove a certain point. The MIT dataset, on the other hand, simply gathers together images that might be worked on professionally, whether dull or interesting, without […]

Read the full article →

Introducing the PPW Tools Panel, version 5

by Dan Margulis June 2, 2018

The Picture Postcard Workflow is more than a decade old, but it never really caught fire until the time of Photoshop CS5, when an imaginative Adobe product called Configurator allowed us to store actions in the form of a panel for easier use. Even in those prehistoric times, PPW was highly automated, so assembling its […]

Read the full article →

When Is There No PPW Advantage? (The MIT 5k dataset 7)

by Dan Margulis March 27, 2018

The arsenal of weapons that we can deploy against recalcitrant images is powerful and impressive. When used prudently, that is. Assuming that because you have them you are guaranteed to create stunning imagery tempts you to try too hard to make it so. When this happens, as we’re about to see, the weapons, including PPW, […]

Read the full article →

How Often is PPW Superior? (The MIT 5k dataset 6)

by Dan Margulis February 12, 2018

Almost any method of correcting images works some of the time. For those interested in PPW, or in taking a four-day class on how it works, the question has to be how much of the time. The MIT study we’ve been looking at offers a unique opportunity to answer that question. It shows the real-world […]

Read the full article →

2018 Applied Color Theory Class Dates Announced

by Dan Margulis January 18, 2018

For those wishing to take color skills to the ultimate level, here are the two dates for Applied Color Theory classes in 2018. • ATLANTA, Wednesday, May 16, through Saturday, May 19. • SAN DIEGO, Wednesday, August 22, through Saturday, August 25. These classes—four long days, limited to eight persons—have changed the lives of many […]

Read the full article →

White Point, Dark Point, Auto Tone: The Simplest Move of All (The MIT 5k Dataset 5)

by Dan Margulis December 10, 2017

“And yet,” I wrote in my first book 25 years ago, “most color correction could be handled by monkeys…a numerical, curve-based approach calling for little artistic judgment…all the advanced techniques are inevitably based on these surpassingly simple ones. The by-the-numbers rules can be stated in a single sentence: Use the full range of available tones […]

Read the full article →