Adirondacks: Comments on individual versions, part 1


Dan Margulis
 

Either I had a lot to say about these submissions or I am suffering from boredom while hiding out from the virus. In view of the length, I’ll cut these comments into two parts.

Comments I made at the top of the same file in Seated in the Grass are equally applicable here:

As with a couple of others, there were so many good versions that nobody should take my quick choices for the par version too seriously, as there are around ten versions that could reasonably have been chosen.

I also suggest that if you are going over this list you download them and compare some versions by toggling back and forth, using Apply Image or similar. Some of the differences are quite subtle and hard to detect otherwise (which doesn’t make them any less useful).

Dan Margulis
*******

1101 We start off with a personal favorite, as well as a favorite of some of the people posting to the main thread.. The person explains:

I sought to create a sense of the wonderful light and overall scene one sometimes finds oneself in after a rainstorm -- dark clouds in the distance, but with a sunny, bright foreground. The car entering from the left reinforces this sense of the scene -- it's coming from a considerably darker area, and so its headlights are on.
I began by setting a neutral on the road. And once again, I made several versions and blended them, aiming for a result that would be an appropriate response to what you, as I interpreted it, wrote about the image: that it was an unusually colorful fall, so represent it as such, but don't overdo it.
In addition to blending a few different versions I adjusted the saturation and luminosity of some parts of the image, masking out others. The sky may stretch the boundaries of what looks "natural", but after trying several versions of the sky I felt I needed those tones of blue to intensify the vivacity of the colors in the trees.

He claims to be concerned about having lost detail in the trees in exchange for his brighter colors, but I don’t see this. There are several versions that have lost detail due to excessive color, #1103 being one, but this one seems OK to me.

According to precedent, since I would rate this as one of the five best versions, it should have been chosen for the par version. However, it is so radically different from the others that it would throw off the blend. So I chose 1110, which is somewhat along the same lines but not as extreme, even though I don’t like it as well as this one.

1102 This pleasant version was made in two passes. The person did the first in a standard PPW way, including MMM. The result is one of the darker treatments of the trees, which looks good but which has the drawback that the shadows at the base of the trees can plug up. He apparently saw this and did a second version in which he used Bigger Hammer to open the shadow areas up. The result is satisfactory but it lacks the punch of the par version.

1103 This version went charging into MMM + CB right after making a few color adjustments. It proved that with these tools, you can make colors as vivid as you like. Unfortunately, in doing so, you can also knock out detail in the colored areas. There’s very little definition in the leaves of the major trees. Also, the red of the most prominent tree is equivalent to the reds found in the flags, which is unrealistic. Still, it’s an attractive presentation, impressionistic, and quite useful for blending in Color mode.

1104 This person also went into MMM and did not like what he saw, feeling that it threw the reds over the top. He therefore cancelled and reran it, this time using only greens as the advisory selection. The idea is successful in that it emphasizes the greens, and also the variation in the foreground orange tree. These things flatter the redness without forcing it to compete with the flags. 

Unlike #1102, this person created some detailing in the sky. That, coupled with the MMM move described above, makes it similar to the par version, which is nevertheless better in spite of the similar coloring. Why? Because if you are going to make an interesting sky you should make its lightest point be roughly equivalent to a normal highlight, if this can be done without damaging the flags. Since the par version does that, its added contrast carries over into the foreground trees.

1105 As in #1104, this person emphasized the reds by means of calling attention to the greens with of a subtle MMM selection, though he used MMM Finetuned rather than MMM proper. The result is a slight green cast in the trees. I think that’s good, I did it intentionally in my own version, because it allows a more flexible portrayal of the reds. This person had another concern:

I think the colors came out nicely, both in the flags the foreground and the background, and even though the background became a bit lighter I think there is enough difference front to back the make the foreground be the eye catcher.

I agree with the conclusion but not the reasoning. Imagine this image without any flagpoles. Then, this guy’s concern would be valid. But the tall U.S. flag is going to focus attention on the foreground trees almost regardless of what we do with color.

1106 The result here is somewhat similar to #1103 in that the colors are attractive but there’s a significant loss of detail in the foreground trees. However, #1103 had better separation of colors, the two main foreground trees are more distinct from one another.

There was a questionable strategy involved in this one. The person immediately boosted saturation in ACR. This may have made it temporarily look better but it was premature. When LAB is available later there’s no big hurry to make the colors bright, better to make that decision later and not risk doing something now that’s hard to reverse.

1107 A number of MMM steps, including channel blending using the green, led to this version in which the reds of the trees become oppressive. This often happens when we are fooling around in LAB, our eyes become desensitized and we don’t see that the colors have become too strong. The solution is to look at the file again a few hours later and blend it with something more conservative if necessary. That way, it would probably be adjusted to something more like #1110. 

There is, however, a more important problem here. I’ve recommended several times over the course of these studies that when you think you are done with the file, run Auto Tone. We hope and expect that this does nothing and should be cancelled immediately. It’s a good two-second insurance policy against a flawed procedure.  Here, Auto Tone makes a distinct improvement.

1108 Chosen for the par version. This one combines the best features of those seen so far. The flags stand out, the two major trees are distinct, subtle detailing in the shadows reminds us of how much redness there is in the site, and the sky is interesting. The person describes his process as “all a bit trial and error rather than a grand plan.” However, the strategy sounds like a grand plan to me: he made one version for color and another for contrast.

1109 As with the previous versions this one presents an attractive scene. It is not quite competitive with either #1108 or the par version because there is not as much color variation; the two trees behind the first three flagpoles seem to merge into one. The reason is similar to that in #1106: boosting color prematurely. The person writes that after trying to get good color in Camera Raw, 

I took it into Photoshop and converted to LAB and tried curves on L, A, and B. All that did was to push large parts of the trees out of gamut and by the time I reduced opacity of each channel to get in gamut there was no point. My submission is a basic global adjustment in ACR.

My suggestion: if you choose to correct color in a raw module the object should be to eliminate obviously wrong color rather than to produce something pleasing in and of itself.  Doing otherwise can pre-empt the use of more powerful tools later.

1110 Chosen for the par version. This version is what #1107 ought to have been; both have chosen relatively dark trees, fair enough, but this one has good color variation and the reds are not oppressive. The person lived in New England for many years and states that these are the colors he recalls. He got them with several PPW steps that happen to make the image darker: blending green into red, Bigger Hammer, false profile/multiply, and H-K, followed by the usual MMM+CB. Note how H-K suppressed some of the color and made it possible for him to be more aggressive once he got to LAB.

Another nice touch: he darkened the shadows in the road, suggesting stronger sunlight and making his brightly-lit foreground more realistic in context.

1111 A first-rate version, emphasizing the foreground by putting deeper and more neutral shadows elsewhere. His procedure was similar to that of #1110, including both Bigger Hammer and H-K. He says he spent about 2.5 hours on this over three separate sessions. The results paid off, I’m not sure I don’t like it better than the par.

To be continued in Part 2.

Join colortheory@groups.io to automatically receive all group messages.