Date   

moderated Re: Case Study: Choir

Dan Margulis
 



On Mar 8, 2021, at 10:24 AM, Dan Margulis via groups.io <dmargulis@...> wrote:

*In the Photos section, 2021 Case Study: Choir
I have uploaded the two original files, as given to the 2009 classes.  There is no zipped sourcefile, no raw file.

A reminder that entries are due in this case study in 47 hours--1 hour earlier than usual, due to the time change in the U.S. tomorrow--at 06:00 eastern daylight time Monday/05:00 Canada/1000Z/11:00 ora italiana

I confirm receipt of entries from the following individuals:

BB
GB
RoB*
FC
HD
JG
DS
JS
KSo
RW
CZ

*indicates that a corrected version was submitted

Entries from the following were at an incorrect size/cropping and would have to be resubmitted:

None


Dan Margulis


Re: King of Beasts: Dan's comments

Robert S Baldassano
 

Dan thanks for the comments.  I think I have been using the auto comparison wrong. I have been creating an auto version from the initial image. And have used that to compare with the final. Should I be creating the auto version from my final version or does it matter.? 

Robert Baldassano


Re: King of Beasts: Dan's comments

Bill Theis
 

I am embarrassed that I did not think to do, as a last step, checking with an autolevel!  (#612)  Stupid, stupid, stupid.  I take my new modified pledge:

Never submit anything any time without a final round of setting endpoints and correcting color.... AND checking with autoTone. 

I've been spending all morning looking at a number of things I correct in my own work... and the autolevel tells me that I somehow drifted into picking a much darker image that is not attractive. 

This is going to pay dividends for all my future work.  Thank you Dan for helping me. 


Dan's comments

Hector Davila
 

Dan's comments are like the Leonard Bernstein of Color Correction.


I don't know if yous have seen Leonard Bernstein conducts West Side Story,

but when the music stops...

everyone waits to see what the Leonard Bernstein of Color Correction has to say.

Because you know Dam Margulis is going to listen to

every musical note on the color photo you just finished playing.





Leonard Bernstein conducts West Side Story

https://www.youtube.com/watch?v=j3SEW63LsaM


(and the Par become the finished record album.)


Hector Davila


Re: King of Beasts: Dan's comments

 

What a powerful and fascinating analysis!! I did the suggested layering in both color and luminosity and can now ascertain where did I go away from the Par. ( #628). This challenge has also taught me that I have to distance myself from my craft as a photographer if I want my work to be closer or at least no that far away from the "optimum" or ideal Par.

That said, it is amazing to notice that the PAR is a hybrid, formally, a non-existing version built from the best treatments, and, at the very least in this specific case- I find the par too soft-contrast for the purposes of printing. Of course, Dan, the expert printer in the room may have a say about this.

Finally, my comment about detaching from the photographic craft comes to the discussion, since I decided to massively darken the 2 upper corners of the image,  as well as the shadow under the tree, instead of looking for shadow detail, easy to pull out, since the darkening in combo with the position of the lion creates a "classical" triangular composition, the lion making an almost perfect diagonal, which helps the viewer stick around the center and upper part of the image and keep browsing around in circles, while the PAR version has no "visual anchors" due to the lack of contrast and the view tends to slip away from the frame. 

Granted, I understand now that this compositional element, contrast considerations for visual impact, deviates from this challenge itself,  more focused in Color and Luminosityso, hopefully, I will not play around with so much extra contrast in the upcoming challenges.  

BTW: I am having quite some fun NOT using sharpening or using just the basic minimum, ( like some clarity during the  DNG process) and that's it.
--
Jorge Parra 
www.jJorgeParraPhotography.com
Miami


Re: King of Beasts: Dan's comments

Dan Margulis
 
Edited

Technique: There won’t be much discussed here because almost nobody did anything out of the ordinary. Many of us luminosity-blended with the green channel, almost everybody used MMM or some variant, and some used one of the Hammer action. Overall color correction used a variety of methods and applications. Many people tried some kind of painting into the lion’s face.
 
The honorary par: I noted in the previous post that there is a “consensus” look that many of us were trying to achieve, and also that there were a dozen or so high-quality entrants. I was a little reluctant to pick five similar images for the par, and instead added two outliers that have strengths that make them suitable for blending, although by themselves they’re not particularly attractive. I realized shortly before posting that we might as well have a second par, an honorary par, derived from five other versions that all follow the consensus model. That one is #609. I personally prefer the “real” par, #636, but #609 is a nice version that got a couple of compliments during the group commentary.
 
The demonstrations: I complained in my first post that too many people had a bad B/A ratio resulting in an unpleasant color feeling. This happens a lot in the outside world when we have to deal with yellowish things that need to seem more golden. Such files can benefit from the LAB procedure that I illustrate with a demonstration at #617. Because it is so generally applicable, I more or less repeat that demonstration in a more complicated way after #621, and have posted an additional image to the Photos folder showing that result.
 
Note on the MIT entrants: Since sharpening wasn’t allowed when the four of these were made, I have sharpened them to make for a fairer comparison to our own work.
 
Dan
*********
 
601 Here’s your nominal opponent, the average result of the five MIT retouchers. I certainly hope you did better than they did.
 
602 Kent Sutorius has already expressed his view that this one is too dark.
 
603 David Remington color-balanced this one in ACR, but not completely successfully. He now comments,
 
 
 I have been thinking that my 603 is a little green. Not drastically so but enough to look off.
 
I was aiming for "realistic" color with good detail and contrast. I still like my image and with a small adjustment to add red to the lion's fur I like it even better. 
 
These remarks came after I complained that it and many other versions had an excessively high B/A ratio in the fur. And I suggest that everyone who’s comparing their own version with the par, or any other version that they happen to like better than their own do the following, for which I use this version and #609 as examples:
 
Demonstration:
*Add a duplicate layer to #603.
*To the duplicate layer, apply #609. Toggle it on and off, and decide on an opinion (Mine: I prefer #609)
*Change the layer mode to Luminosity, toggle the layer on and off, and decide again (My opinion: I prefer the lighter #609)
*Change the layer mode to Color, toggle the layer on and off, and decide again (#603 looks green by comparison, so I don’t need to state my opinion, because it’s going to be the same as everybody else’s.)
 
Going through this process can be useful because it can pinpoint the issue. I’m not saying that #603 is a bad version, quite the opposite, it’s probably better than half of the ones in the folder, and it’s surely a lot better than #601. But given a choice between #603 and #603 with #609’s color added the judgment is going to be unanimous, even though the difference is small in the overall scheme of things.
 
One solution is to follow the procedure outlined in the demonstrations after #617 and #621
 
604 Chosen for the par version, measuring a nice brown in the fur, and one of the redder manes.
 
605 Too dark. Wrong color, the reddish mane measures magenta. Also needs Auto Tone. 
 
606 This person’s view:
 
 
Preference call between neutralizing the warm light, and harnessing it.
 
Harnessed indeed. The color is approximately as good as that of the par. The luminosity is another story. It looks washed out, with no detail in the white fur on the chin. This is the opposite of #603 because the par version can improve luminosity, but not color.
 
607 John Furnes laments:
 

I obviously overdid ”get rid of the cyan”-thing, and it became more magenta and yellow than perhaps a lion is.
However I was content at the time, and even though my third attempt was more like the par is, I chose No.1
I need to learn more about the Blend If possibility as I tried to reduce the yellow patch to the right of the animal, but gave it up as more than the patch were affected.
 
I’d say that this is a prelude to versions #617 and #621, where a more golden look is desired. The problem with this one is not so much the color of the animal is that this color could only be gotten in fairly strong sunlight. Granted that, the version is too dark, although it would be fine if the lion were more neutral.
 
608 John Gillespie laments:
 
 
I have achieved a bright, sunny look - if you assume that the beast has been sent on loan to the Zoological Gardens here in London. In December.
 
The colour and weight of the par are very nice, although it is perhaps a little soft for what is after all an apex predator (albeit a notoriously indolent one). Mixing in a little of the luminosity from one of the harsher versions (such as my own) improves it I think.
 
That this version is less saturated doesn’t correct the color imbalance. It needs remedial action along the lines of what I’ll show in #617 and #621.
 
Meanwhile, this is an example of how a certain defect can blind us to the inherent strengths of an image. I again suggest the Luminosity/Color test as in the demonstration after #603. This time, according to me, the results are similar but more decisive. With the par version on top of this one in Luminosity mode, I think #608 as it stands is preferable—better than the par.
 
When the layer mode is changed to Color, however, not only is the par’s superiority even more blatant than it was in the #603 comparison—the result as it stands, with the par’s color and this one’s luminosity, is very striking indeed, possibly the best version we’ve yet seen.
 
609 This is the honorary par version. Hard to dislike it.
 
610 Almost blurry, lacking contrast in fur.
 
611 This is my 2017 three-minute version. I anticipated (correctly, as #601 confirms) that it would be a lot better than what the MIT retouchers could produce, because of the impact of channel blending and the MMM script. However, my light fur is green and the animal’s color is about as bad as #603. Another illustration as to why we need comparisons. At the time, I really thought this was a hot version. And compared to the MIT versions, it certainly was, but it falls kind of flat today.
 
612 Bill Theis had at this with the usual suite of PPW tools. On a luminosity layer, he blended the green into the red, Darken mode. He used Velvet Hammer. He used a false profile and, of course, MMM + CB. For these reasons he left all the MIT entrants in the dust. Not the rest of us, alas. That’s what happens when you do all the right things and forget to check the work with Auto Tone at the end.
 
613 Nice texture, but like many others, imbalanced toward yellow/green. There are what seem to be highlight and shadow samplers that offer clues as to why. A sampler in the lightest fur in the cheek and it measures (1) 17. In the dark fur at the left of the animal, 7a16b. I’d prefer to choose the interior of a nostril as the shadow point, it’s more likely to be black, whereas this dark fur could be a dark brown. If it is a dark brown, then the desired relationship is A=B. Having the B that much higher confirms the yellow-green cast.
 
614 Cartoonish color. Many people went to great lengths to eliminate any trace of blue in the background log. I don’t think it’s necessary to be as doctinaire as that, but it surely can’t be this blue. Meanwhile, the light fur in the chin measures (4)a(7)b when it would need to be, if not white, then a yellow of come kind. The fur in the mane measures 37a 39b, basically pure red, hardly tending toward yellow at all. 
 
615 Chosen for the honorary par. This person apparently missed out on the channel blending so that its luminosity falls short of that of the par. However, careful handling of H/S sliders and some local work made the color excellent, possibly better than the par.
 
616 Chosen for the par version. Throughout these studies people have often posted saying that they liked their own version, when my own view was something far different. This is the first time I can recall the opposite happening. Saïd Nuseibeh says, after reviewing the rest of our work,
 
I wasn't thinking Happy Hour as many others here were, just dappled light. I pulled the cyan out of the body and, especially, the face but left it in the environment... reading it as open shade bouncing a blue sky.
 
Post-Bellagio, I did enjoy pitching the warmth of the lion's mane against a contrasting cooler surround.
 
With the benefit of hindsight and this august assembly, I was clearly too damn judicious because the cyan sticks out like pee on snow. 
 
Setting aside this editorial commentary, Saïd decided early that attention needed to be drawn to the face, so he implemented a radial gradient mask and went to work. The result is that the face, by the standards of other good versions, is quite normal. The reddish mane measures 56L25a35b, which is essentially the same as the 58L26a40b of #609, the honorary par version.
 
The two are different in two very significant ways. #609 looks much like any of its parents. It is a consensus of skilled people as to what this picture should look like. As such, it’s impossible to dislike it. Many people, OTOH, would say they dislike #616 because there are so many cool colors where they obviously don’t belong. However, in the age of color grading, they really draw attention to the lion’s face, without forcing us to drive more color into that face. This makes it an excellent blend-source candidate. I would never say this was acceptable as is, but as part of a par ensemble, it serves its purpose.
 
617 Paco Márquez:
 
 
I went for the "golden hour" type of mood. After seeing the Par version, I now see I put too much gold into the "golden hour" mood.
 
I was very happy with mine, seeing it in isolation but once I compared it to the Par I can see how far from the "true" light of the original I ended up with.
 
I still like mine but not as much as I like the Par version.
 
The theory here was fine, the practice left something to be desired. Getting the “golden hour” look is a good thing. The “yellow hour” look is something else, and that’s what’s been had here.
 
Of course, there is no agreed-upon definition of gold. My suggestion is something like 80L20a70b. IOW, I say that the AB values are both positive, the B should be quite high, and between 3 and 4 times as much as the A. 
 
The animal's fur is no mirror. The golden light will change its natural color somewhat, but not a tremendous amount. In “normal” lighting conditions, that fur would be a light brown. In the flank area that I’ve been measuring, that suggests, as discussed that the B should be not more than twice as high as the A. Now if golden light is hitting it then the ratio gets bigger. Maybe the B should be two-and-a-half times the A in that case, but three times the A would be excessive.
 
But what do we have here? The yellowish detail-free patch behind the lion measures 86L4a69b. The B is more than 15 times the A. This is a yellow, not a gold. And in the fur, my sample reads 69L8a28b, the B being 3.5x the A. That’s gone way too far toward yellow. So we either need to lower the B or raise the A, or both. The problem with lowering the B is that the image will seem less colorful, so I don’t like that idea. Instead,
 
Demonstration
*Convert #617 to LAB.
*Add a duplicate layer.
*Select the A channel, and blend the B into it, Overlay mode, 100%. This is going to add a lot of magenta wherever there is a strong yellow component. Normally we’d have to worry about green being added to blue as well, but inasmuch as there isn’t any blue in this picture we can forget about that part.
*Look at the composite now. The effect is intentionally wildly excessive, so that we can be sure we’re getting the correct parts of the animal.
*Reduce opacity of this layer to taste. My taste is about 25%.
*Since the greenery also has a strong yellow component, magenta will be added to it, which will desaturate it. We have to decide whether we want that. It could be right in some cases but here, my decision is that it’s a bad idea.
*Therefore go to Blend If, the Underlying Layer slider, and exclude everything that is negative in the A channel. That restores the grass and background without affecting the lion, which is always A-positive.
*Now toggle the layer off and on, to see the difference between a Golden Hour and a Yellow Hour.
Alternatively, go to the newly posted #637 to see what a similar move produced when applied to #621.
 
618 Chosen for the honorary par. Edward Bateman before:
 
 
Some of my goals were to have no blue in the logs and good texture in the lion.
 
And after:
 
 
One of the things I wanted to accomplish was a greater sense of spatial depth by darkening the foreground and background. Cool (and dark) colors often seem to somewhat recede while warm (and light) colors seem to somewhat advance (or so I believe). So by darkening especially the foreground, I wanted to put more visual focus on the lion. While at first glance, my lion seems a bit dark (and perhaps it is), it is only about (an average) 5-6 points (measured in Lab) darker than the par. So I still think the par looks a bit spatially flatter than mine. A quick check shows that I like about 40% of par blended into mine.
 
I also used a small bit of an old darkroom trick - lightning the face just a small bit to put more visual focus on the lion’s face. One thing I often look at: I’ll squint at the image to see contrast - is my primary visual focus the area with strongest (within reason) contrast? (Since eyes seem to focus on areas with the best contrast). In this sense, I like mine better than the par- the lion face has better contrast when compared to the overall image in the par.
 
No quarrel from me, but my question is why there was no attempt to check quality with Auto Tone after completion. I don’t know that Auto Tone is an improvement but it certainly would lead to using Auto Contrast, which is an improvement.
 
619 Fur too yellow, but the big problem is the image is too flat. Running Auto Tone makes a huge improvement. So we’ve now seen four examples: this one. #605, #618, and #612. What is the big problem with always checking Auto Tone after the job is believed complete?
 
620 Chosen for the honorary par. A bit darker than some of the others, but plenty acceptable. Note the positive impact of this guy’s decision to dodge the eyes, which helped counter the sense of darkness.
 
621 Gerald Bakker says in retrospect:
 
 
I missed at least one thing: the more saturated, orange manes. In the original the color difference is clearly visible, but somewhere during my processing (probably blend-if on a color boost layer) this got lost. Well, always something to learn. 
 
I think this is a typical PPW image. There is a strong blue cast that must be fixed. A green blend works well to bring out detail. Color boost as the original is way too dull. I didn't do not much local correction this time. As Dan says, it's one of the easier exercises. 
 
I agree completely with the second paragraph. As for the first, I agree that more color variation toward red in the mane would be helpful but I believe this can be made into a serious competitor to the par without it. This version takes the same approach as Paco’s #617 in that both go for more color and less neutrality in the animal’s coat. And (although they are very different in the way they treat contrast) they share the same problem: instead of a warmer, more golden color, they measure too yellow. In my sampling point in the animal’s flank, I recommended no greater than a 2.5/1 B/A ratio, and probably more like a 2/1. That, incidentally, is what the par has: 7a14b. Granted that this one seems naturally yellower, we can increase this ratio somewhat, but #621 as it stands has it 9a36b, or 4/1. The fix is similar to that for #617 but more complicated.
 
Demonstration:
1) Open #621 and add a duplicate layer.
2) I assume that the par version is not available to assis us. As Gerald notes, the green channel has the most contrast for blending purposes, but not quite enough IMHO. So, just do it again: apply the current green channel to the RGB, but use Darken mode, as otherwise parts of the greenery may get lighter.
3) Change mode to Luminosity. I left opacity at 100% but it can be adjusted downward to taste.
4) Move the file into LAB, without flattening. Check to see whether the appearance changes and, if so, whether you’d prefer to flatten in LAB or RGB. My conclusion is that there’s no significant difference.
5) Add a composite layer.
6) As with #617, select the A channel, and apply the B to it, Overlay mode.
7) As with #617, having verified that the blend has worked, reduce the opacity to taste. I chose 25%, just as with #617.
8) This makes the lion more golden, but it also desaturates the grass, and we have to decide whether to accept that. If not, we exclude it with Blend If. In #617 I did not want to accept it, but in this somewhat softer treatment I think it’s beneficial to play down the grass. So, I take no action here.
9) Another composite layer, this time to darken the greens. To do so, apply the B, inverted, to the L in Overlay mode.
10) As this move has darkened the lion as well as the grass, go to Blend If and exclude anything that is A-positive, since the greenery is the only thing that’s A-negative. Doesn’t make a difference whether the This Layer or Underlying Layer slider is chosen.
11) Option-click the slider near the midpoint to break it into halves, and spread them slightly so there isn’t a harsh transition.
12) Reduce the opacity of this grass-darkening layer to taste. I chose 40%.
13) Yet another composite layer.
14) Apply the B, not inverted, to the LAB in Overlay mode. Most everything now becomes brighter and much more colorful.
15) Add a black layer mask, thus temporarily nullifying move 14.
16) Set a brush to white, and paint over the eyes on the mask, so that the eyes in the final result will be lighter and more colorful.
 
At this point I would convert to RGB and declare victory, saying that it’s now at least as good as the par. The checkpoint in the fur is now 18a36b, a fine ratio, and more than twice as colorful as the par.
 
I have posted the result as #637 in the Photos folder.
 
622 Too light, there was no channel blending. 
 
623 MIT Retoucher C achieved reasonable detail. Here, however, the fur is actually neutral—it measures 1a(1)b at my sample point.
 
624 John Castronovo:
 
My goal was to bring attention to the eyes, nose and whiskers. I purposely subdued the grass and the bright yellow in the background to help bring attention to the face and mouth. Other than that, I wanted a natural look with nothing else exaggerated so that the viewer would gravitate to the lion's facial expression of contentment about lounging in the warm sun. After seeing the par, I guess I might've gone for more contrast and a brighter look, but I think the greenery is too distracting in the par. I liked 609 for a snappier result and I believe it works very well when blended in luminosity mode with mine. 
#609, as we now know, is the honorary par.
 
625 Equal A and B in the fur produce the hue you would expect if this were a dark brown horse. For a brown this light the fur has to be distinctly yellower, higher in the A. That’s why this version looks so weird.
 
626 Chosen for the par version, proving that two wrongs sometimes make a right. I find it overdone. I see it as more of an auxiliary version than a standalone entrant. Once having chosen the bluish #616 for the par, however, I needed a strongly yellow treatment to balance it, the obvious candidates being #617, #621, and this one, which happens to be my least favorite of the three. However, it also happens to make the best blending partner because of its powerful sharpening.
 
627 Chosen for the par version. I think this is one of the two best “consensus” treatments of this image. The  person comments, wisely, that
 
 
This was a real color fatigue test—so much yellow! It’s rough judging color while being so overwhelmed.
 
His correction was done with non-Adobe products so I can’t say much about specifics. He did note that he lightened the eyes, painted down the saturation int he yellow area to the right of the animal, and killed a couple of catchlights.
 
628 Something bizarre about the lighting here. How can the yellow background be so brilliant, when there are what seem to be shadows all over the lion? Fur measures magenta. This person mixed up a brew of Smart Objects to produce it
 
629 Chosen for the honorary par. Good softness, Good fur color, but a general lack of variation, shown by the lack of redness in the mane. This apparently was deliberate, as the notes state that the objective was to make the background more realistic in the sense of being similar to the lion’s own color. This person spent a lot of effort making the log less blue, leaving it magenta instead. See note to #631.
 
630 In the course of some complicated maneuvering, this person lost control of the color. Like #629, it’s too yellow in the reddish parts of the mane. The darker fur in the body is OK, but the lighter fur in the belly and leg, which should be the yellowest as well as the lightest part of the animal, are magenta.
 
631 MIT Retoucher E’s version is similar to #629 in feeling, but too light. Three in a row, now, with lack of decent color in the mane.
 
632 Chosen for the honorary par. The string of three colorless manes in a row comes to a violent end with my own entry, which is closer to the par version than any of the par’s actual parents. I actually prefer it to the par, largely because this one illustrates how important it is to make piercing eyes.
 
633 A pleasant version if you want something that dark. Looks like a spotlight on the cat due to painting in clarity and texture into the face.
 
634 Chosen for the par version. Fine color values. Painted color boost into the lion.
 
635 This one comes from students of Edward Bateman. They join the ranks of those needing the technique in the demos of #617 and #621.
 
636 The par version. If you want an alternative, I offer three: 1) #609, the honorary par; 2) #608’s luminosity with the color of #636, for a very dramatic effect; 3) #637, which I posted as the result of my demonstration after #621. Blend the three together to taste, but even if you just average the three it looks to me much better than the par. In fact, I’ve just posted that version as #638.
 
 
 


Re: King of Beasts: Dan's comments

sj_90000@...
 

To add to Dan’s insights. There’s also the benefit that the green channel more closely mimics the percieved brightness of a scene. Both the red and blue channels can be quite jarring/contrasty depending on the content.
 
HTH – Steve
 
Date: Thursday, March 11, 2021 07:47 AM
Subject: Re: [colortheory] King of Beasts: Dan's comments
 
 

On Mar 11, 2021, at 1:40 AM, Christophe Potworowski <christophe.potworowski@...> wrote:
 
Novice question: Why Green into Red? Why not Blue into Red?
 
The character of the individual channels may vary depending on the method of acquisition. Some people did blend with the blue as source, which is better than no blend at all. The biggest reasons to avoid the blue in this particular case are first, it would darken the eyes, when we all would want to lighten them instead, and because it would pre-empt certain decisions. Either channel would bring more detail into the lion but the blue would also darken the greenery. That may be something we want to do, but there’s no need to lock ourself into that decision now.
 
Also, as a general rule: if the choice is between using the blue or the green as a source, unless the blue is obviously more suitable, use the green. The green channel is always less noisy than the blue, and we don’t need to add that noise elsewhere.
 
Dan


Re: King of Beasts: Dan's comments

Dan Margulis
 



On Mar 11, 2021, at 5:34 AM, Gerald Bakker <gc.bakker@...> wrote:

My #621 is in both lists, and indeed it's more orange-yellow than most other versions. The main motivation to deliver something so yellow is the color of the logs behind the lion. I think these can be neutral or slightly B-positive. In the par, they measure B -10 or even stronger blue, which to me looks incorrect. Of course, these logs are unimportant image elements, but on the other hand, their blueness should be an indication that the whole image is too blue.
I added the par on top of my version in Color mode, and honestly, I don't find it an improvement. 

Me either, but that’s only half of the test I suggested, which was to put the par version on a top layer, and evaluate whether it has made an obvious and undeniable improvement overall 1) when the layer is set to Luminosity; 2) when it is set to Color.

Applying that test to David Remington’s #603 and Gerald’s #621 yields the following results, according to me.

#603
with par on a Luminosity layer: the par has a lighter overall feel and would likely but not necessarily be preferred.
with par on a Color layer: obvious and undeniable improvement.

#621
with par on a Luminosity layer: obvious and undeniable improvement.
with par on a Color layer: no clear improvement, possibly even a slight loss.

How do we account for this difference, considering that my post accused both David and Gerald of the same sin, having the B/A ratio too high?

In Sunset on the Beach especially, and Bellagio and Concert on the Beach somewhat, there were many possible interpretations and we saw examples of most of them. In this lion image there is far more of a consensus on what we’re looking for. The best representative of this is probably not the par but rather #609; I’ll explain later. If #609 is considered the reference, then we can easily pick out 10 or maybe 12 entrants that are similar.

Just because it’s the consensus view doesn’t mean it’s the correct one. #621, #617, and #626, possibly others, have a different agenda. They want a considerably warmer feel.

#603 OTOH is pretty much a “consensus” correction. Put, say, #609 on top of it in Color mode, and it’s an obvious improvement, because artistically the two versions are trying to achieve the same things but the color of #603 is Wrong.

The same test with #617 or #621 fails, as it should. If the strategy calls for a color different than the consensus, why should anyone be surprised that imposing consensus color doesn’t improve matters?

The fact that aspirin doesn’t prevent Covid doesn’t mean it’s not a useful medication. The fact that par color doesn’t improve #617 or #621 doesn’t mean that they shouldn’t  comply with certain recommendations. The color of #617 and #621 is just as Wrong as that of #603, there just isn’t at present an easy comparison that would prove it. I will therefore produce one in the next post.

Dan


Re: King of Beasts: Dan's comments

Dan Margulis
 



On Mar 11, 2021, at 1:40 AM, Christophe Potworowski <christophe.potworowski@...> wrote:

Novice question: Why Green into Red? Why not Blue into Red?

The character of the individual channels may vary depending on the method of acquisition. Some people did blend with the blue as source, which is better than no blend at all. The biggest reasons to avoid the blue in this particular case are first, it would darken the eyes, when we all would want to lighten them instead, and because it would pre-empt certain decisions. Either channel would bring more detail into the lion but the blue would also darken the greenery. That may be something we want to do, but there’s no need to lock ourself into that decision now.

Also, as a general rule: if the choice is between using the blue or the green as a source, unless the blue is obviously more suitable, use the green. The green channel is always less noisy than the blue, and we don’t need to add that noise elsewhere.

Dan


Re: King of Beasts: Dan's comments

Gerald Bakker
 

On Wed, Mar 10, 2021 at 10:49 PM, Dan Margulis wrote:
I happen to have measured the same points in every version. I also measured the highlight areas, but we didn’t seem to have many problems there, so no need for discussion. With respect to the lion’s body, I have a list of all those who have a B/A ratio of 3/1 or higher. Also, a list of those having the mane at significantly greater than 2/1. In the first category there are way too many guilty parties, namely ##603, 605, 608, 610, 611, 612, 613, 617,  618, 619, 621, 623, 624, 626, and 635. On the second count, the following are indicted: ##601, 603, 621, 623, 629, 630, and 631. And on the other side, the following have it as essentially a pure red: ##605, 614, and 628.
 
What are these people guilty of, if the B is too high? Another way of putting it would be, the A is too low. Four other LAB-based descriptions might apply: too yellow, not magenta enough, not blue enough—or too green. Those last two words, if true, explain why only two of the above-named versions made it even into my top ten, which I picked without looking at any numbers. Maybe you can get away with such slight inaccuracies in a more difficult image like the three previous studies, but where almost everybody is getting something halfway decent the error is noticeable.
My #621 is in both lists, and indeed it's more orange-yellow than most other versions. The main motivation to deliver something so yellow is the color of the logs behind the lion. I think these can be neutral or slightly B-positive. In the par, they measure B -10 or even stronger blue, which to me looks incorrect. Of course, these logs are unimportant image elements, but on the other hand, their blueness should be an indication that the whole image is too blue.
I added the par on top of my version in Color mode, and honestly, I don't find it an improvement.
 --
Gerald Bakker
https://geraldbakker.nl


Re: King of Beasts: Dan's comments

Christophe Potworowski
 

Novice question: Why Green into Red? Why not Blue into Red?

Christophe


Re: King of Beasts: Dan's comments

David Remington
 
Edited

Great comments from Dan based on experience and observation.
 
I'm going to jump in ahead of the chop with a mea culpa. I have been thinking that my 603 is a little green. Not drastically so but enough to look off.
 
I was aiming for "realistic" color with good detail and contrast. I still like my image and with a small adjustment to add red to the lion's fur I like it even better.
 
I find the par version a bit too red however and I prefer the contrast in my version. Layered with mine in luminance mode switching back and forth I find the par a bit washed out.
 
After reading Dan's initial comments I went looking for lion color and there is was in an archived version of the ISCC-NBS Dictionary of Color Names. "Lion" NBS-ISCC color numbers 76 and 77.
 
sRGB 196, 154, 116 and 136, 102, 72 for light and dark versions. Lab 67L 13a 26b and 46L 11a 23b. Right on Dan's 1:2 numbers.
 
https://web.archive.org/web/20121122220617/http://tx4.us/nbs/nbs-l.htm
 
https://www.munsellcolourscienceforpainters.com/ISCCNBS/ISCCNBSSystem.html
 
I made a new version for myself using this ratio as a guide and it is an improvement. I also lit up the eyes a bit. This has been a great exercise. I enjoy seeing what the group comes up with and thinking through Dan's critiques. I look forward to reading the details to come on this one.
 
David


King of Beasts: Dan's comments

Dan Margulis
 
Edited

This exercise is the equivalent of the Monument Valley study of last year, in the sense that it’s the one image out of the ten that most clearly shows the advantages of PPW. Indeed, this is very like the canyon images that are commonly used to promote work in LAB: lots ofsubtle colors that need to get more variation, yet we can’t afford to make them too lurid.
 
There are not as many traps to fall into as in our last three case studies. Furthermore, the highlight and shadow points are obvious. For all these reasons, I consider this the easiest exercise of the ten, with the possible exception of Hotel Lobby. Like that one, here there are many blending possibilities but one (green into red, Luminosity mode) stands out and was adopted by many of us. Also, the MMM script, which is crucial for getting variation into the lion, was generally used. For these reasons we have a dozen or more decent entries that could reasonably be chosen for the par.
 
We expect, also, that we should do much better than the MIT retouchers, who presumably used none of these tools. And this is indeed the case. Their average result happens to be #601. Most of us did significantly better, and better also than their two best individuals, who are #623 and #631.
 
Yet a lot of our work that is obviously better than #601 doesn’t stand up so well next to the par. There is no disgrace is preferring the par to your own version; in our studies typically only one or two people, if that many, can honestly say they consider their version to be as good as or (occasionally) better than the par. And if we do have a slight preference for the par, it’s usually because the par tends to have more accurate color, not so much better contrast.
 
If you’re mystified as to why the par is better, here’s a valuable test. Put the par as a layer on top of your version. Toggle it on and off. Now, change its mode to Luminosity, so that what you see is your color, with the par’s contrast. Again, toggle between the two. Then, change the layer mode to Color, so that you see your contrast, its color. And toggle again.
 
The normal result is that you will decide that the par is slightly better in both respects. Occasionally, though, you will find that the two versions are about equal in one respect, but in the other, the par is grossly superior. We saw and discussed an example of that in the Mantillas exercise where one person actually had better color than the par’s, yet his contrast was much worse. I regret to say there are around half a dozen such instances in this set. That shouldn’t be happening.
 
If the par’s luminosity is much better than yours the explanation is probably simple. You probably didn’t do the appropriate channel blending.
 
If the par’s color is much better than yours the explanation is likely more complicated, and in a way that is applicable to other similar images. Start by asking yourself: what is the basic color of this animal? I think we can agree that it is neither lemon-yellow nor tomato-red, but something between those two extremes. Which of these is the best description?
*Reddish yellow
*Yellowish brown
*Brown
 
What are the desired values, then, for a typical patch of darker fur, nearly neutral, in the top half of the animal’s flank?
 
Brown is a species of red, although much less saturated than a tomato. If we measure LAB values, in principle we should get something like A=B with both positive. That, however, can only be quite a dark brown, such as a walnut table, or a dark brown horse. As the brown gets lighter, it also needs a higher B value. So, a light wood like beech, or a lion for that matter, can have the B twice as high as the A. Going any further than that technically makes it a reddish yellow rather than a yellowish brown.
 
What happens if you ignore the need for a higher B, and have A=B? Well, then you get something like #625. The hue is perfectly fine for a dark brown animal, but it doesn’t work in a lion. Or, in another way, you get #614, which has a much better body. The mane admittedly needs to be redder than that. But it can’t go all the way to A=B.
 
I have measured a representative point in the body, and also in the part of the mane that we expect to be somewhat redder. What would acceptable values be for each area?
 
We should acknowledge that the lion is being hit by sunlight that could push it more toward yellow. So for the body, I think I’d accept anything up to about B=A*2.5. For the mane, which we know is redder, I’d say that anything worse than B=2A is definitely wrong.
 
I happen to have measured the same points in every version. I also measured the highlight areas, but we didn’t seem to have many problems there, so no need for discussion. With respect to the lion’s body, I have a list of all those who have a B/A ratio of 3/1 or higher. Also, a list of those having the mane at significantly greater than 2/1. In the first category there are way too many guilty parties, namely ##603, 605, 608, 610, 611, 612, 613, 617,  618, 619, 621, 623, 624, 626, and 635. On the second count, the following are indicted: ##601, 603, 621, 623, 629, 630, and 631. And on the other side, the following have it as essentially a pure red: ##605, 614, and 628.
 
What are these people guilty of, if the B is too high? Another way of putting it would be, the A is too low. Four other LAB-based descriptions might apply: too yellow, not magenta enough, not blue enough—or too green. Those last two words, if true, explain why only two of the above-named versions made it even into my top ten, which I picked without looking at any numbers. Maybe you can get away with such slight inaccuracies in a more difficult image like the three previous studies, but where almost everybody is getting something halfway decent the error is noticeable.
 
These images likely fall in that unpleasant category described above, where the par is markedly superior in terms of color. Translation: this is the end of a three-part series in which coolness vs. warmth is a key issue. In Beach at Sunset, those who engineered warmth into the scene were successful. In Bellagio, it was the reverse. And here? You be the judge.
 
Because a problem with these same colors often arises in other contexts, I’ll have a pair of demonstrations on how to rectify it with a simple LAB blend tomorrow when I post comments on individual versions.
 
A couple of lesser points.
 
First, it’s very helpful to lighten the eyes. A lot of us did so and it makes the animal seem more real.
 
As for desaturating the red nose, or any of the greenery, I’d say it’s personal preference. Some of us were willing to let the background have some color, others tried to suppress its saturation, or even make it darker. A big controversy was what to do with the yellow area in the background, Some think it suggests sunniness, even if unrealistic. Some say that lions are designed to blend into the environment. They therefore lessened, or retouched it out. OTOH, this particular lion wasn’t given a choice of background environment. So I am fine leaving that yellow in. It’s a matter of taste, as is leaving the log behind the animal slightly blue. In principle it’s a by-the-numbers violation; in practice, we live in an age of color grading and that blue helps bring out the subtleties in the yellowish-brown (repeat, yellowish-brown) of the lion.
 
Dan


Re: King of Beasts: Results

Robert S Baldassano
 

 

 

Sent from Mail for Windows 10

My entry was 619. Compared to PAR 636, it appears my colors were close to par but the differed in L with par being much brighter. While I can see that my version could be brighter, I am not sure if I would like it as bright as par, perhaps blending with par at about 50% to 70% in luminosity. I had commented to Dan that I had created 3 versions the other 2 were much darker, and that if I was in a hurry the auto version might even fit the bill. It seems  so far my entries are all a bit to dark. I do think I am slowly improving.

 

Robert S Baldassano


Re: King of Beasts: Results

John Gillespie
 

Mine is 608.

I have achieved a bright, sunny look - if you assume that the beast has been sent on loan to the Zoological Gardens here in London. In December.

The colour and weight of the par are very nice, although it is perhaps a little soft for what is after all an apex predator (albeit a notoriously indolent one). Mixing in a little of the luminosity from one of the harsher versions (such as my own) improves it I think.





.


Re: King of Beasts: Results

Edward Bateman
 

Hello everyone - been really enjoying this - and studying more with each challenge.

Mine was number 618. Some of my goals were to have no blue in the logs and good texture in the lion.

One of the things I wanted to accomplish was a greater sense of spatial depth by darkening the foreground and background. Cool (and dark) colors often seem to somewhat recede while warm (and light) colors seem to somewhat advance (or so I believe). So by darkening especially the foreground, I wanted to put more visual focus on the lion. While at first glance, my lion seems a bit dark (and perhaps it is), it is only about (an average) 5-6 points (measured in Lab) darker than the par. So I still think the par looks a bit spatially flatter than mine. A quick check shows that I like about 40% of par blended into mine.

I also used a small bit of an old darkroom trick - lightning the face just a small bit to put more visual focus on the lion’s face. One thing I often look at: I’ll squint at the image to see contrast - is my primary visual focus the area with strongest (within reason) contrast? (Since eyes seem to focus on areas with the best contrast). In this sense, I like mine better than the par- the lion face has better contrast when compared to the overall image in the par.

And just FYI - number 635 was created by a small group of students in one of my classes. Dan was kind enough to allow its submission, and they were very interested today to see how they efforts stacked up! I personally don’t know any other university level teachers who teach any of Dan’s techniques, aside from me. And I have to say, I’m proud of what they are doing - we all recognize that their critical color skills and ability to see more precisely has already made huge strides this semester. They themselves are rather surprised at how different they are seeing things. So thank you for letting them play along! And I too really appreciate the challenge from so many smart people! Thanks Dan!

-Edward Bateman


Re: King of Beasts: Results

Paco
 

Mine is 617. I went for the "golden hour" type of mood. After seeing the Par version, I now see I put too much gold into the "golden hour" mood.

I was very happy with mine, seeing it in isolation but once I compared it to the Par I can see how far from the "true" light of the original I ended up with.

I still like mine but not as much as I like the Par version.

All the best!

Paco


Re: King of Beasts: Results

John Furnes
 

Mine is 607

 

I obviously overdid ”get rid of the cyan”-thing, and it became more magenta and yellow than perhaps a lion is.

However I was content at the time, and even though my third attempt was more like the par is, I chose No.1

I need to learn more about the Blend If possibility as I tried to reduce the yellow patch to the right of the animal, but gave it up as more than the patch were affected.

 

So..

 

John Furnes

 

 


Re: King of Beasts: Results

Said Nuseibeh
 

Well, mine is 616. I wasn't thinking Happy Hour as many others here were, just dappled light. I pulled the cyan out of the body and, especially, the face but left it in the environment... reading it as open shade bouncing a blue sky.

Post-Bellagio, I did enjoy pitching the warmth of the lion's mane against a contrasting cooler surround.

With the benefit of hindsight and this august assembly, I was clearly too damn judicious because the cyan sticks out like pee on snow. Oh, and I neglected to review my endpoints before converting to JPEG.


Hang down my head Tom Dooley


Re: King of Beasts: Results

john c.
 

Mine is 624 and my goal was to bring attention to the eyes, nose and whiskers. I purposely subdued the grass and the bright yellow in the background to help bring attention to the face and mouth. Other than that, I wanted a natural look with nothing else exaggerated so that the viewer would gravitate to the lion's facial expression of contentment about lounging in the warm sun. After seeing the par, I guess I might've gone for more contrast and a brighter look, but I think the greenery is too distracting in the par. I liked 609 for a snappier result and I believe it works very well when blended in luminosity mode with mine. 

301 - 320 of 33205