Topics

Case Studies in Retrospect


Dan Margulis
 

I would like to have a thread on what we have learned in this series, but first a couple of commercial plugs.

If you think that this type of image studies or competitions or whatever you may call them is useful, be aware that Daniele Di Stanio runs a site that specializes in them. At the end of each study he presents a high-quality video, which may be a little easier to absorb than my dense written commentary. It is a pay-to-play site, because running such studies takes a lot of work, I'm here to tell you.

Second, the Chevreul/Margulis book On the Law of Simultaneous Contrast of Colors ought to be on your reading list. It has no mention of Photoshop as such but much of it has to do with the philosophy of how to present art, including my comments on when stuff like MMM is especially useful. And if you've read it already, please help me out by posting a review on amazon.com.

The reason I bring it up is that on several occasions people remarked that their own version was much worse than the par. A couple of times I responded that in that case, they might consider going back to the original and starting over, trying to duplicate that par image, which I thought would be harder than it looks. Those who tried it said that it was pretty easy, if we know that the par is the target.

My question, then, is if you feel that the par is better than yours, and you know how to produce the par on demand, then why didn't you produce it in the first place? That's where Chevreul can be helpful.

More later in the day.
Dan


Jim Sanderson
 

Dan, I am currently in Anchorage AK  I will be without internet service for the next couple of weeks.  I wanted to thank you (and all the other participants) for your time and excellent instructive comments. I’ll have to reread your newest book and rework many of the exercises again when I return home. 

Thanks again for all the insights and ideas. 

Jim Sanderson 

, with purpose and great thought. 

On Aug 7, 2020, at 5:32 AM, Dan Margulis via groups.io <dmargulis@...> wrote:

I would like to have a thread on what we have learned in this series, but first a couple of commercial plugs.

If you think that this type of image studies or competitions or whatever you may call them is useful, be aware that Daniele Di Stanio runs a site that specializes in them. At the end of each study he presents a high-quality video, which may be a little easier to absorb than my dense written commentary. It is a pay-to-play site, because running such studies takes a lot of work, I'm here to tell you.

Second, the Chevreul/Margulis book On the Law of Simultaneous Contrast of Colors ought to be on your reading list. It has no mention of Photoshop as such but much of it has to do with the philosophy of how to present art, including my comments on when stuff like MMM is especially useful. And if you've read it already, please help me out by posting a review on amazon.com.

The reason I bring it up is that on several occasions people remarked that their own version was much worse than the par. A couple of times I responded that in that case, they might consider going back to the original and starting over, trying to duplicate that par image, which I thought would be harder than it looks. Those who tried it said that it was pretty easy, if we know that the par is the target.

My question, then, is if you feel that the par is better than yours, and you know how to produce the par on demand, then why didn't you produce it in the first place? That's where Chevreul can be helpful.

More later in the day.
Dan


Dan Margulis
 

Now that these case studies are drawing to a close I want to thank all those who included notes of encouragement and/or statements of progress in the summaries sent to me of the procedures they followed. I did not have time in recent weeks to respond to comments not directly pertaining to the images, but I’m going to be reviewing everything and will reply as needed.

But, speaking of review, it would nice to hear from the group about what was accomplished, what worked, and what didn’t. I ask this because my experience in reading evaluations over the 25 years that I taught Applied Color Theory is that students and instructors often don’t see eye to eye with the instructors (or even themselves) about which images were the most useful. My suspicion is that people think the most useful image is the one they screwed up, rather than the one that made the best points.

Alas, we have to doubt that the ACT course can ever be held again. Still, by force of habit, I’d repeat one of the questions on the evaluation form: name three or more images that you found particularly useful. Those that, over time, don’t get their share of votes, get thrown out and replaced by others for future classes. And those that get lots of votes get kept even if I think there are better options.

These series started with a problem presented by Roberto Tartaglione. By popular demand we continued, but I chose the next ten in advance, trying for a good diversity of challenges and subjects. Unless somebody wants to argue about it, and granted we are dealing with only 11 and not 110 images, I’m satisfied with the distribution. Some are easy, some are hard, some are ideal for PPW and others not, some call for artistic judgment and others are obvious, some have major lighting issues and others are fine. And in this day and age, I say it is appropriate to include something shot with a smartphone and also a scan of an elderly color print.

That’s not to say that some of these examples aren’t instructive enough to be particularly good representations of their class. Hence the need for comments. Tom Hurd made a good start with his listings of the things that he says he has learned over the course of the 11 exercises. I hope there will be more. (A listing of the case studies we worked on is appended.)

I won’t name my own favorites just yet, but here’s an example of what I’m looking for when I review these exercises: I think an unusual number of our images were able to make good use of the H-K action. OTOH it’s under-used, presumably because people don’t understand when and why it’s useful. I’ve never seen as good a demonstration/explanation as the one in Cinque Terre: take #501, a version from one of the MIT retouchers. Its colors are not wrong, but they are dull and boring.

Assume you have decided to correct to add more color, and that your tool of choice is Adjustments: Hue/Saturation. First, run H-K, and then do your color boosting and save the file for later comparison. Now, go back to his original, without the H-K, and see if you can get as good a result with H/S only. I don’t like to say “impossible” but certainly “very difficult” would serve.

I am glad that most of us found the series enjoyable and instructive.

Dan

***********
The 11 case studies, in order:
Carnival
Veiled Bride
Niagara Spray
Cinque Terre
Colosseum
Panama 1978
Monument Valley
Toast to Greece
Seated in Grass
Adirondacks
Red Rose


Hector Davila
 

I used the H-K for the very first time on these test because I needed something that does what H-K does. I cannot imagine anything out there that does what H-K does, but I was able to get the
effect I needed with it.

I ran H--k once at 100%, then I ran it again at 100% and reduce to 75% and vola, just the effect I needed. I guess I was just lucky.

Hector Davila

On 8/8/2020 4:26 AM, Dan Margulis via groups.io wrote:
Assume you have decided to correct to add more color, and that your tool of choice is Adjustments: Hue/Saturation. First, run H-K, and then do your color boosting and save the file for later comparison. Now, go back to his original, without the H-K, and see if you can get as good a result with H/S only. I don’t like to say “impossible” but certainly “very difficult” would serve.


Alec Dann
 

Dan,

I've really enjoyed the case studies.

I encourage you to think about converting Advanced Color Training to an online course using a platform like Teachable.  There participants can go through the course materials and work through the example images but they also have the opportunity to ask questions about specific sections and get answers in a day or two.  If you don't have time for answering the questions, I'm sure you have some advanced students who could step in as teaching assistants.

Alternatively, you could do a paid course where a limited number of students submit images that use the PPW workflow and you do a live or recorded meeting where you discuss improvements or alternative process paths to consider.

All said, I'd love to get more of this.  I only participated with two of the case studies because I've been reading Modern Photoshop Color Workflow and decided I needed more of that first.

PWW has given me a path to enhancing color that offers a lot more control than the methods I've been using and I'm excited to fully understand it.

Alec


k_d@...
 

In all I do, I rarely pick favorites; so none picked by me.
In my mind I think I learned something from every image; regardless of how good my result was. Of course nearly all images present new challenges; so the more the merrier.

First is by looking at all the other images and seeing how some had a different approach that is an idea to use somewhere else too.

Second is the analysis by Dan and others of any of the images; from what seems wrong, what is the goal, what is an approach, and what went wrong.

Thirdly by Dan's analysis of what is tricky and important and how to try various methods to reach the target for the image.

And lastly, a continuing message, is if you have the time, it is probably always better to blend several attempts...par images are almost always the best overall. As I proceed now I plan for a "fork in the road" and to use at least 2 results....sometimes more.

Thanks again Dan for all your books, PPW, and your efforts to help so many of us. The case studies add to the wealth of 'learning by doing' and being measured/evaluated; thus knowing if I am learning and making progress.

Doug Schafer


John Gillespie
 

First, many thanks to Dan for being so generous with his time and expertise throughout these challenges, and for providing us with entertainment and education in these strange times.
Also thanks to all the participants for sharing your ideas and approaches.

I first came across Dan's books a few years ago, and the ideas in them have really improved my images, but in isolation it is not always easy to verify that the techniques are being applied correctly or to the best effect. So the opportunity to get feedback directly from Dan, and to compare my efforts with others has been of great value and a genuine privilege. 

To that end all of the exercises have been of use to me, but if I were to pick the three that I found most interesting and enjoyable I would say:
Toast to Greece
Panama 1978
Colosseum

Another fascinating aspect of the process has been when someone has produced a quite different take on an image - that has been a bit of an eye-opener in terms of considering the possibilities for how a scene could be interpreted.

In terms of specific learnings:

- Comparing and evaluating different versions of the same image is quite hard (for me anyway) - and a vivid example of the law of successive contrast. Often the changes in the second image are perceived so strongly that they can be mistaken for errors, and the order of evaluation can have a strong effect on the ranking.
- Judging how much colour/saturation to add to an image is harder than determining the optimal contrast in my experience, but H/K can allow you to be more adventurous with saturation - Dan's suggestion re the Cinque Terre image is a good validation of this I think.
- Separating issues of matters of taste vs. objective correctness is much easier when you can see other participants' results.
- Finally, the whole process has reinforced quite clearly Dan's ongoing recommendation to make use of multiple versions. In almost all of the challenges you could select any two images at random, blend them in some way and get a version that was better than both of the originals - the blend is your friend.

I hope that at some point there will be opportunities for further challenges.


Dan Margulis
 



On Aug 9, 2020, at 4:48 PM, k_d@... wrote (and John Gillespie similarly):


And lastly, a continuing message, is if you have the time, it is probably always better to blend several attempts...par images are almost always the best overall. As I proceed now I plan for a "fork in the road" and to use at least 2 results....sometimes more.

I would not have predicted that the pars would do quite that well, but I agree that they did. Perhaps I should have expected it.

Compare any two versions of a given image at random, and ask yourself whether a significant majority, say two-thirds, would prefer one version to another. I’m experienced at asking myself that question and then verifying my opinion with others. Without seeing these two versions in advance, I’d have to predict that the chances of one being voted better are about 40%, since I imagine that 40% of the time the jury would prefer the other and 20% of the time it would have no preference.

Now suppose that we compare one of these versions to five randomly chosen ones. What are the odds of it winning against all five? This is impossible to state accurately, because it’s not like flipping a coin five times, where what happened on the first four flips has no bearing on the result of the fifth. But each time our version wins it becomes more likely to win against its next opponent. If it has already won against four opponents it has shown itself to be good enough to be a heavy favorite against the fifth. So, I’m going to speculate that the odds of being better than all five randomly chosen opponents are slightly less than one in ten, let’s be generous and say 10%.

For the MIT study I took 150 random images, each corrected by five retouchers. In each case I averaged the five, then compared the average to each of the five parents. According to my scoresheet the odds of the average being better than an individual parent were not 40% but 74%. The odds of being better than all five were not 10% but 24%.

I was thinking that our par versions would show somewhat similar results but this was stupid. Of the five MIT retouchers often one or two would do a notably poor job, and this would detract from the average version. Someone who did a good job, therefore, had a potential advantage over the average. In our case this was not true: the par versions weren’t created from random efforts but rather from good ones, the best five, say, out of 25.

And the results were convincing. I didn’t actually count these up like I did with the MIT study but my impression that in at least half of these studies the par would have been voted better than any of its five parents. There were instances where people came up with something at least competitive to if not better than the par, but they were rare.

I seem to recall posting once that blending multiple versions is probably most advantageous to the inexperienced, because averaging may cover up certain deficiencies in one version. But these studies suggest that maybe blending is even more advantageous to the experienced professional.

Dan


Robert Wheeler
 

When the Carnival image was posted, I had just purchased Dan’s Modern Photoshop Color Workflow book and discovered the Color Theory discussion group via the link in the book’s online resources. Because my learning from the image challenges has been inseparable from my concurrent reading, I have difficulty identifying three images as most helpful in themselves.

 

I experimented with the Carnival image but did not submit my result. My main learning was confirmation that I would need to learn much more to be effective at the extremes of color correction. I also discovered that members of the group are fluent in a language that I would need to learn. I knew about channels but did not know how to “blend” them. I knew about CYMK and LAB color spaces but did not understand how to use them well. I was unfamiliar with many commands in the Photoshop “Image” menu, which had been cast as a forbidden “destructive editing” zone by many of my earlier learning resources. It was not easy to find clear and comprehensive explanations of how to use “apply image.” My response was to accelerate my reading of Dan’s book and to get used copies of two older books about channels. Photoshop Channel Chops helped a lot, and The Photoshop Channels Book had a few good nuggets despite the author’s annoying attempt to sound folksy.

 

The Veiled Bride image was my first attempt to deliberately follow the steps of a PPW workflow: initial evaluation and color correction followed by contrast adjustment and then revisit color adjustments. I had not yet become familiar with the PPW panel.

 

The Niagara Spray image was my first practical experience using apply image for channel blending. Practice with this and with images downloaded from Dan’s book resource site provided useful anchors for what can seem abstract and mysterious if just reading about it. I made an action to make and save all the channels (R, G, B, L, a, b, C, Y, M, K) into the initial image channels panel for quick comparison and planning.

 

For the Cinque Terre image, I began to experiment with the actions triggered by the PPW Panel. Finishing up Dan’s book and reading the panel PDFs would take some weeks. I found the sub-panel display of examples common LAB values helpful as I began to consciously measure a* and b* values as part of image evaluation. This was also my first venture into making versions by different methods then blending versions together. I began the multiple-week project of reading the recently released On the Law of Simultaneous Color Contrast.

 

With the Colosseum image, Dan’s detailed analysis of L values across three points in all the images opened my eyes to the value of a more systematic approach to detailed image analysis than I would have attempted on my own. Very instructive and useful.

 

The Panama 1978 image tossed in the issue of scanner artifacts to address. Looking for resources led me to trying Pattern Suppressor 2.6 with useful results (but not in time for my challenge image).

 

For the Monument Valley image, I revisited key parts of Photoshop LAB Color edition 1 which I had read years earlier with less than perfect comprehension. (Version 2 is essentially unavailable: out of print and used copies asking $200 to $600! Will save up for Version 3 if it ever emerges.) I used some of the PPW Panel actions, some manual curves in LAB, and blended versions at the end. Pleased to see my progress validated a little with my entry selected as a component of the PAR version.

 

In the Toast to Greece image, the mixed lighting presented a useful challenge. This one was not easy. I see in my original notes that I did include the H-K action as part of the process, although omitted that from my posted summary. Learned a lot from working on this one, and again pleased to see my entry picked as part of the PAR version.

 

The Seated in Grass image led me to create the “most colorful” entry via using more parts of the PPW Panel actions than earlier, with my notes admitting to a gradual reduction in timidity about color resulting from participation in these challenges. I am beginning to understand more about the actions, and was happy to see my entry selected as part of the PAR version.

 

I found the Adirondacks image exceptionally useful for seeing how differently others approached the correction.

 

The Red Rose image provided a good reminder not to confuse color with structure when figuring out what to “correct.”

 

Despite my reservations about mixing up images and books as sources of learning, perhaps it is fair to say the Toast to Greece, Seated in Grass, and Adirondacks images provided strong lessons for me.

 

I have learned more than I can summarize, and gradual integration into evaluation and processing of my own images has already begun. Dan has served as a good model for the potential benefits of re-processing the same image over time. I expect to revisit these challenge images and re-read the PPW panel PDFs over the next several months to get a firmer grasp on the methods and when to use them.

 

Deep thanks and appreciation to Dan and to all participants for accelerating my learning via these image challenges.

 

Robert Wheeler


Harvey Nagai
 

First and foremost I would like to extend profuse thanks to Dan for conducting this series
of case studies.  It was a golden opportunity for me and his generosity with his time and
commentary made it more than worthwhile.

Of course I would like to thank the group as a whole, your entries and commentary were
vital to the experience.

====

Useful images:

"My suspicion is that people think the most useful image is the one they screwed up,
rather than the one that made the best points."

By that criterion 10 of 11 images proved too useful for me.

Generally I would say the most useful images were the ones which had the most diverse
results.  These offered more examples of good interpretations and successful techniques
with plenty of contrasting examples to show why some well-considered decisions turned out
suboptimally or even disastrously.

Personally the most useful images were the five which involved skin tones, even though they
frustrated me the most.  Most of my focus was and will be on ancient family photos, and
my modus operandi has been to wrestle faces into something recognizeable then move on to
something more gratifying.  I should be able to do better by them.

The image I found the least "fulfilling" was the "Red Rose" because after a trip into
my backyard where a "Tropicana Hybrid Tea Rose" has thrived despite my brown thumb,
I'm wondering if the more correct interpretation of the color was a lighter orange/coral.
(In which case the M.I.T. par image would have the most accurate color, albeit with vague
contrast.)  But this is a narrow personal perspective, the case study was an interesting
bookend to the "Carnival" case study.

The image I found most gratifying was the "Adirondacks" image, it was an image in which
that fell into my comfort zone and my entry showed that.  But more importantly, being
in that comfort zone allowed me to (finally) study what the H-K actions could do for
the image for both the background as a whole and the shadows in the foreground.  So call
it a good "beginner's" image for H-K, at least for me.

====

What was accomplished:

My personal objective/hope for these case studies was to gain a better understanding
of how to "read" images, the "what" and the "why".  The "how" I can figure out.

To be honest I'm not sure how much I learned, but I went from wondering what to do
beyond the obvious (Veiled Bride, Niagara Spray, Cinque Terre) to having a bit more
of a global vision (A Toast to Greece, Adirondacks, Red Rose).  From not particularly
noteworthy to a positive object lesson, a par candidate, and a background comparison,
my last entries at least got Dan's attention, so there's that, and I'll gladly take it.

My entries: 217, 313, 420, 506, 610, 724, 822, 920, 1018, 1114, 1209

====

What worked:

Blending, for the later case studies.  Evolving from "stupid/par" blends of your own
work is not easy, intentionally making an image specifically for blending is hard
(unintentional monstrosities like my "A Toast to Greece" entry don't really count,
do they?).  But I managed a somewhat effective one for my "Adirondacks" entry.

H-K, also for the later case studies.  Early on I used the action without really
understanding what it was intended to do, I either liked what I saw and kept it
or I didn't like what it did and undid it, in either case moving on after giving
the action just one shot.  It wasn't until the "Seated in the Grass" case study that
I tried to edit the image with an intention to use H-K.  For the "Adirondacks" case
study I dissected both actions then based edits on what I expected them to do.

====

What didn't work:

"Stupid blending", for the earlier case studies.  It took awhile to learn that
if you blend images as your interpretation evolves, you're de-evolving your
latest variant back towards what you've already decided to discard.  If you
have a firm but flawed interpretation, all you are doing is reinforcing the
same flaws.  Usually the result is a bit better, but it's not using blending
to its full potential.

(Note: I wrote this before Dan posted his recent response about blending,
but I think my experiences are a valid addendum: there are stupid ways to
defeat the goodness of "stupid blending".)

BTW, the group pars I've done for my own curiosity are impressively good,
which made it all the more bamboozling that blends of my own variants were not
much better than any single one, until I thought about why.

My comments on the results of the first few case studies didn't add anything
to the conversations, so I stopped.  My images spoke for themselves and how
they differed from the better images, and I didn't do anything extraordinary
that was worthy of explanation.  If they deserved commentary Dan and other
group members made the cogent points.  If there are case studies in the future
hopefully I can contribute usefully to the discourse.

====

Follow-up:

Eleven case studies in eleven weeks was a tough haul, especially if you have
a sisyphean bent.  The end means there is much to review, weeks of it.

So right now a month subscription to weekly Color Duels case studies seems too
daunting.  Practising on random images from the M.I.T. FiveK dataset is a
practical alternative, albeit without any feedback.

For these case studies I purposely put Dan's books aside and avoided googling
articles and tips, it was just me and my brain, an often sorry partnership.
Now it's time to hit the books, not the least of which is Dan's latest.

====


James Gray
 

Dan,
I really appreciated you doing this.  As others have noted the exercises were quite like your ACT course.  I learned a lot even though I did not create multiple versions in most cases or actually follow up with additional work after seeing what other participants had done.  I think the 3 image I found most useful were Cinque Terre, Monument Valley, and Toast to Greece.  The first two because they more closely resemble the kind of images I work on most often.  I am not sure I can put my finger on why I thought Toast to Greece was so useful.  I put a lot of work into it.  One thing that amazed me was the wide variation in the outcomes of the various participants.  One thing I learned that may seem trivial in comparison to many comments -- In the 2nd step of the PPW when blending channels in a layer to be converted to luminosity mode many other channels and blending modes may work than what I was using.  For example, using multiply mode to blend green into red or vice versa may give good results.  Sometimes overlay may work with something other than the A and B channels.  Another issue was to watch the color of hair and skin.  I found the Niagara Spray to be frustrating because I thought that somehow the background should be made more distinct and detailed.  I could not and neither could anyone else.  I found out that I seem to have a tendency to just go through the same steps for almost all images.  I need to be more flexible in my approach. In some ways I wish that we could have gotten more feedback during the work on one image.  I did trade images with one other participant before he or I submitted our final images.  That was useful.

Jim Gray


On Sat, Aug 8, 2020 at 5:27 AM Dan Margulis via groups.io <dmargulis=aol.com@groups.io> wrote:
Now that these case studies are drawing to a close I want to thank all those who included notes of encouragement and/or statements of progress in the summaries sent to me of the procedures they followed. I did not have time in recent weeks to respond to comments not directly pertaining to the images, but I’m going to be reviewing everything and will reply as needed.

But, speaking of review, it would nice to hear from the group about what was accomplished, what worked, and what didn’t. I ask this because my experience in reading evaluations over the 25 years that I taught Applied Color Theory is that students and instructors often don’t see eye to eye with the instructors (or even themselves) about which images were the most useful. My suspicion is that people think the most useful image is the one they screwed up, rather than the one that made the best points.

Alas, we have to doubt that the ACT course can ever be held again. Still, by force of habit, I’d repeat one of the questions on the evaluation form: name three or more images that you found particularly useful. Those that, over time, don’t get their share of votes, get thrown out and replaced by others for future classes. And those that get lots of votes get kept even if I think there are better options.

These series started with a problem presented by Roberto Tartaglione. By popular demand we continued, but I chose the next ten in advance, trying for a good diversity of challenges and subjects. Unless somebody wants to argue about it, and granted we are dealing with only 11 and not 110 images, I’m satisfied with the distribution. Some are easy, some are hard, some are ideal for PPW and others not, some call for artistic judgment and others are obvious, some have major lighting issues and others are fine. And in this day and age, I say it is appropriate to include something shot with a smartphone and also a scan of an elderly color print.

That’s not to say that some of these examples aren’t instructive enough to be particularly good representations of their class. Hence the need for comments. Tom Hurd made a good start with his listings of the things that he says he has learned over the course of the 11 exercises. I hope there will be more. (A listing of the case studies we worked on is appended.)

I won’t name my own favorites just yet, but here’s an example of what I’m looking for when I review these exercises: I think an unusual number of our images were able to make good use of the H-K action. OTOH it’s under-used, presumably because people don’t understand when and why it’s useful. I’ve never seen as good a demonstration/explanation as the one in Cinque Terre: take #501, a version from one of the MIT retouchers. Its colors are not wrong, but they are dull and boring.

Assume you have decided to correct to add more color, and that your tool of choice is Adjustments: Hue/Saturation. First, run H-K, and then do your color boosting and save the file for later comparison. Now, go back to his original, without the H-K, and see if you can get as good a result with H/S only. I don’t like to say “impossible” but certainly “very difficult” would serve.

I am glad that most of us found the series enjoyable and instructive.

Dan

***********
The 11 case studies, in order:
Carnival
Veiled Bride
Niagara Spray
Cinque Terre
Colosseum
Panama 1978
Monument Valley
Toast to Greece
Seated in Grass
Adirondacks
Red Rose


Dan Margulis
 



On Aug 10, 2020, at 3:14 PM, Harvey Nagai via groups.io <hnagai@...> wrote:

Generally I would say the most useful images were the ones which had the most diverse 
results.  These offered more examples of good interpretations and successful techniques 
with plenty of contrasting examples to show why some well-considered decisions turned out 
suboptimally or even disastrously.

This is the best summing up of the experience, though I certainly thank Doug, John, Robert, and Jim for their comprehensive responses.

When I picked this set I was looking for such possible variation, so I either went with things I thought were subject to differing creative treatement, or else ones that required good technical skills (since I knew that some people would show us what happens when certain important steps are missing or mishandled).

Personally the most useful images were the five which involved skin tones, even though they 
frustrated me the most.  Most of my focus was and will be on ancient family photos, and 
my modus operandi has been to wrestle faces into something recognizeable then move on to 
something more gratifying.  I should be able to do better by them.

Fair enough, you concentrate on certain things and others on others. However, in this day and age there is less specialization. Anybody who knows this field can expect to be asked to correct almost anything. Yesterday, I got a whole slew of such ancient images dumped on me, so I’m sympathetic to the above. I had some offline criticism of including Panama 1978 in the set, and I disagree. In your case, it’s theoretically the most significant, just as based on what I know of Jim’s work Monument Valley may be the theoretically most significant for him, and anybody who feels they shouldn’t work on such images can go jump in the lake AFAIC.

The image I found most gratifying was the "Adirondacks" image, it was an image in which 
that fell into my comfort zone and my entry showed that.  But more importantly, being 
in that comfort zone allowed me to (finally) study what the H-K actions could do for 
the image for both the background as a whole and the shadows in the foreground.  So call 
it a good "beginner's" image for H-K, at least for me.

I agree now, but I didn’t when I first put the set together. None of the first seven images was especially easy, and then came Toast to Greece, which is the hardest of all IMHO. So to avoid potential suicides among group members I decided to finish with three easier ones. Red Rose is maybe not easy for the general public but given that we had gone through Carnival we should have done OK.

I’m not sure of the value of Seated in the Grass except to reinforce some of the earlier skintone images. Adirondacks I imagined was going to be kind of a throwaway, and it wasn’t, I’d definitely use it again.

I note that there are several votes both on and offline in favor of both Adirondacks and Toast to Greece, so let me point out an interesting difference. It’s hard to make something really ugly out of the Adirondacks shot; the question is how attractive it can be made. A Toast to Greece, OTOH, is full of invitations to disaster, which a number of us accepted.

I’m slowly re-reviewing everything and have only got through the first three, I’ll have more to say later. It looks to me like Veiled Bride called for a lot of technical skill, it seemed like the best images were from the people I know to be the best and most experienced practitioners. Cinque Terre, which has those technical issues but calls for more creativity, is going to be on my own list for one of the best exercises. Niagara Spray also had a little of both. I expect that when I’m through with the re-evaluation, I’m going to say that the group as a whole had the poorest results on that one.

Comments are great so far.

Dan


James Gray
 

DM:  However, in this day and age there is less specialization. Anybody who knows this field can expect to be asked to correct almost anything. Yesterday, I got a whole slew of such ancient images dumped on me, so I’m sympathetic to the above. I had some offline criticism of including Panama 1978 in the set, and I disagree. In your case, it’s theoretically the most significant, just as based on what I know of Jim’s work Monument Valley may be the theoretically most significant for him, and anybody who feels they shouldn’t work on such images can go jump in the lake AFAIC.

JG:  I fully agree that Panama 1978 was a good exercise even though it did not make my top 3.  I certainly have had to work on images that were scans of very old prints.  I certainly learned from my mistakes on that one.  It was also valuable to get some feedback on what to do with the texture of the print.  I gather the answer is -- try to decrease the texture, but do not try to get rid of it totally.  I do not understand why anyone would complain about the selection when participation was purely voluntary.  Anybody who feels that some image was not something they should work on should skip it.

Jim Gray


ROBIN MARK D'ROZARIO
 

Dear Dan,

Thank you very much for running these exercises. It has been a good opportunity to practice and learn. Also a welcome distraction during the  Covid-19 lockdown.
Thank you as well to the other members of  the forum for their critiques and the techniques used on their submissions.

The three images I found the most challenging were:

1. Carnival - I had never edited an image in CMYK before. Apart from the restrictions associated with working in an unfamiliar workspace I got so focussed on the red robe that I turned the flesh yellow.
2. Niagra Spray - the image was full of traps and artifacts, thanks to being already heavily processed, and required a degree of restraint to avoid aggravating these.
3.Cinque Terre - this required judgement on how to add believable colour variation and contrast. I so fell in love with the colours I got from MMM that I pushed them to unbelievability and neglected contrast.
I will add an unsolicited No.4. A Toast to Greece- a very tricky night shot  with skin , plenty of colour - reds/blues/greens and white - all under crossing casts along with a wide range in exposure.

On the ten images post Carnival I went through my notes and checked which PPW actions were used most.

MMM and CB - 7 submissions
BH, FP/Multiply and H-K - 5 submissions
Skin Desaturation - 3 submissions
S/H - 2 submissions
LH-1 submission

So H-K featured in 50% of my RGB submissions. Certainly more than I have used it while practicing with the PPW in the past. I also used it on a discarded version of the Carnival image and it provided a great deal of contrast to that version.

Blending images proved indispensable - 9 out of 11 submissions. Some of my better submissions and indeed some of these images, like Niagra, Colosseum and A toast... really can't be done justice to without several versions being blended. On the other hand I still need to discipline myself to make more divergent variations as opposed to versions with slight differences.

Some observations on my submissions:
1. I did better with the two night shots Colosseum and Toast...Perhaps because there is greater allowance for personal interpretation in these images.
2.The three landscapes were among my worst submissions. I struggled to balance realistic colour and contrast on Cinque, Monument Valley and Adirondacks. Here my inexperience showed.
3. Images with a dominant central subject like Veiled Bride and Red Rose were led to better submissions. Images with a background that also needed carful handling, Niagra and Seated..,  led to either the background or foreground subject suffering.
4. Getting skin colour right or at least better was challenging for me in Carnival, Veiled... , Niagra .. and Seated..

As Gerald Bakker and others have already mentioned, falling in love with ones work is fraught with danger. I have repeatedly become enamoured with the colour and  contrast that I produced or fixated on some particular idea while processing these exercises.And then proceeded to marvel at my work and neglect other aspects of the image! The greater the love the worse the result has been, by itself and in comparison to others. Working in isolation it's quite a challenge to step back and question whether ones ideas are sound or that there may be a better result to be had by doing things differently.

My first introduction to your writings was with PPW, book and panel, in 2018. A very advanced point for a beginner to land on. I found it too difficult and only returned to it in earnest about a year ago. Since then I have worked my way backwards through Photoshop Lab and your instructional videos, on the Ledet and Kelby training sites. I've learnt more in the past year than the previous ten but these exercises have exposed the huge gaps in my knowledge and understanding that still exist. Day before yesterday I bought an electronic version of Professional Photoshop to help remedy this.

The Chevruel book is currently undeliverable to India. Whether this is due to policy or the current Covid restrictions for courier services is unclear.

Once again thank you and best regards,

Robin Mark D'Rozario













Roberto Tartaglione
 

Thank you Dan for commenting on dozens of images and summing up the best strategies for each exercise. Thanks to all those who have kept me company during these hard times, I learned a lot from all of you.
I will soon answer your question about my preferred list of exercises, but if you wanted to ask me which thread would you like to explore in the future, my answer would be the CMYK world.

Maybe in these days, there is less need for this skill: books and flyers are less and less printed but on the other hand editors and agencies ( I speak for my context only) rely less on prepress companies and often they ask photographers to provide CMYK files if they do not want the batch conversion from INDESIGN to PDF. I think a prepress expert is irreplaceable but sadly in many areas, 
they are no more involved in press jobs.

The chapters on Professional Photoshop 5° devoted to this topic are 
essential and always valid, but maybe more case studies could be of interest to more people. It could be an idea for a 6° edition (I do not consider MPCW a 6°ed…)

That said, my list is:
1) Carnival ( for the reasons above )
2) Cinque Terre: unfortunately when you are on assignment you can not always choose the right moment for the shooting. Managing color and contrast can help to save your neck in some situations.
3) Red Rose: I really underestimated this image and the comparison
with many entrants, showed me how far you can go with the rights steps.

Three important points common to all cases that I hope to fix in my mind:
a) fixing the right endpoints (Yes we all know, but sometimes the choice
doesn't work  to the best)
b) starting from a flat original is often, but we can say almost  always the best choice. I've convinced myself of that during these "challenges".
Despite your recommendations(Modern Photoshop Color Workflow, ch4)
I always tried to achieve the best in the RAW converter software; It is not a wrong choice if you have not much time to spend on an image, 
but if you want to reach a better level, I agree that a flat image gives you more room for manoeuvre.
c) compare and create a "Par" version was not in my workflow but it should be definitely at least in any case of doubt.

°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°°
At the end of my letter, I wish to give my answer to your question:
"My question, then, is if you feel that the par is better than yours, and you know how to produce the par on demand, then why didn't you produce it in the first place? That's where Chevreul can be helpful.

"On the law of the simultaneous contrast…" is not an easy reading, I need to restart a study of it, but what I learned is, in extreme simplicity, that there are colours and shapes that can fit together, others do not.
To figure it in advance is another story.

Thanks again for this experience,
Let’s hope for  the best for our future!

Kind regards
Roberto





Dan Margulis
 



On Aug 13, 2020, at 12:22 PM, Roberto Tartaglione <roberto@...> wrote:


Thanks again for this experience,
Let’s hope for  the best for our future!

Having finished a careful review of all these exercises, here’s my view of the series, after which I’ll respond to some of the individual posts.

I reiterate that I think the eleven represented a good mix of topics. If we had been doing twenty, then there would have been an architectural shot, a wildlife study, a shot of several people, and a product shot. All of these topics were well represented in the MIT study, but I thought there were higher priorities if I only could choose ten.

All six of the MIT images had reasonable color out of the box; I wish I had picked one or more that didn’t. As it stands it wasn’t likely that anybody could truly screw up one of the MIT studies, or the Carnival or Adirondacks shot for that matter. Niagara Spray, Panama 1978, and Toast to Greece could and did blow up in people’s faces.

Since we had so few where accurate color was an issue it should not be a surprise that we had a large number where control of detail—contrast—was critical. It was possible to have a good version of Veiled Bride and/or Seated in the Grass without really having first-rate contrast. The other nine, however, did demand it. I’d say that in six of these nine some form of channel blending was needed to be competitive.

This suggests that the H-K action was more commonly needed in this series than it would be in real life. It’s generally useful where 1) there are a lot of colors of approximately the same saturation, as in Cinque Terre and Adirondacks, or 2) there is one large strongly colored object, such as Carnival or Red Rose.

OTOH the MMM action, which introduces color variation even at the cost of changing hue, wasn’t featured as much in this series as I think would be the case in real life. It was very valuable in Cinque Terre, Monument Valley, and Adirondacks, but not so much anywhere else.

The Hammer actions were valuable in the fleshtone images and Niagara Spray.

Including the best work of the MIT retouchers should have boosted our spirits because they weren’t competitive with our best work except in Colosseum. But this is not surprising. They don’t know channel blending, so Red Rose was beyond them. Without PPW, they couldn’t touch us in Monument Valley and Cinque Terre. And without the fleshtone control offered by the Hammers and/or channel blending they could not manage Veiled Bride or Seated in the Grass, although I’d have to say that they were having a bad day on Veiled Bride and did worse than expected.

When did *we* do worse than expected? Niagara Spray, I’d say. We did *better* than expected, IMHO, on Monument Valley and Adirondacks.

Toast to Greece was the most difficult, followed by Panama 1978.

Creativity was called for in Niagara Spray, Cinque Terre, Colosseum, Toast to Greece, and Adirondacks. The pure technicians did better on the others. Red Rose, for example, either you know how to handle this kind of red or you don’t—not much room for difference about the objective.

It was valuable that we had such an open discussion of when it is permissible (desirable) to depart from reality, usually in the direction of stronger color. We almost all showed a willingness to be unrealistic with Cinque Terre, Monument Valley, and Adirondacks, and in the backgrounds of Niagara Spray and Seated in the Grass. It’s important to understand that just because unnaturally bright colors are good in such images it would be a big mistake to try the same treatment with, say, Veiled Bride.

I know this for a fact, because I fell into that trap. I chose my 100 images randomly from the 5,000 in the MIT set. By coincidence, one of my early picks was from the same shoot as Veiled Bride. In this one, the groom was carrying the bride up the steps to the same brick building found in our case study.  I must say that I did an excellent job of bringing out detail in the bricks, and was considerably irritated when I compared, and lost, to the MIT retoucher average, which demonstrated that interesting bricks detract from the subject.

So when Veiled Bride itself turned up fifty images later, I knew better than to repeat my mistake. That’s how we learn. I hope others have had the same experience in this set.

My own picks for most instructive images are Veiled Bride, Cinque Terre, and Adirondacks.

Dan






Paco
 

 Hello to all and thanks to all!

For me it has been a great learning experience and a welcome distraction. I am not a retoucher per se because I only work on the images I shoot commercially and over which I have as much control as possible when lighting them. But getting a review from Dan is worth participating. Very daunting and for my ego, a nerve wracking and sometimes painful experience.

We who participated faced the challenge. We all did well sometimes, and crashed and burned in others.

I learned the most from the ones that were not very successful because I feel that the best lessons come from those. They show where the minefields are and how to avoid them. On the other hand, there were some results which were "terrible" but I personally liked very much and can see how, with the right tweeks one can make good use of the "wrong" way of doing things.

Having participated and exposing oneself to criticism is not that pleasurable and it takes courage, so to those that feel they did not do well, kudos for diving in! One learns the most from the mistakes we all make. I remember Dan telling me at the end of a class in  Atlanta, when reviewing the results of the final day "after the crap you've lately been handing in, today yours are some of the best in the group." Coming from him those were sweet sounding words.

I learned that besides the weight of good technique, personal taste is also a large part of how one's work is perceived. That someone does or does not "like" what I came up with is valid, but always a debatable opinion.

If one wants to get better at this game, groups like Daniele Di Stanio's offer an amazing learning experience. I was lucky to have been in a class with him and know Daniele has great command of technique and is also a great teacher. Give his group a try!

Of course, having done these exercises with Dan as a teacher is a privilege which I hope will be repeated again one day.

Carnival for me was a hard one but very enjoyable because there was a time where as a professional photographer everything that you handed in had to be CMYK. Believe me, having to explain to an inexperienced Art Director/Client why that beautiful, shiny and bright RGB file, on paper and CMYK inks looked "wrong," made having found Dan Margulis an advantage over my competition which has kept me workin still. I think of Dan every day that I sit on my computer to work on my photos and how much I owe him for his teachings. I can tell you I've met others from around the world who feel the exact same privilege.

So, thanks Dan! You really are a special man! Il più grande maestro!


k_d@...
 

Our thanks again for everything and helpful final words.

And questions:
I think most have learned how to channel blend via apply image to channels and use for image correction....it is a valuable tool as you pointed out in your summary.
I'm thinking we can achieve the same result using channel mixer adjustment layer in place of apply image channel blending.....it seems to be the same thing using a different Ps tool/technique.

And the questions are:
Will channel mixer work OK? and why did you not use it in preference to channel blending via apply image?
How can we use channel mixer to get your channel blending results; i.e. what are the basic procedural steps and perhaps a sample image done both ways?

I'm thinking it will work but I have not been successful doing so, yet. (I have not given up)

Doug Schafer


Thomas Hurd,MD
 

I will say that I have tried channel mixer, but image>apply image is first of all much easier to begin with. 

I have been using Channel Power Tools, which lends itself beautifully to apply image. 

I think that working on the channels for blending and masking is one of the most wonderful thing about Photoshop in image correction.

When I had no idea what I was doing in Carnival, having never produced a CMYK image before, I took the RGB channels and essentially copied them into CMYK. From there I used adjustments like curves, etc. but also the brush tools, like dodge and burn, and the brush itself to paint in shadows and highlights for the individual channels. I Emma’s able to get a better shape to the colored image that way.

I didn’t use the brush techniques on the channels frequently in the other exercises, except during masking.

Tom Hurd

On Aug 14, 2020, at 2:56 PM, k_d@... wrote:

Our thanks again for everything and helpful final words.

And questions:
I think most have learned how to channel blend via apply image to channels and use for image correction....it is a valuable tool as you pointed out in your summary.
I'm thinking we can achieve the same result using channel mixer adjustment layer in place of apply image channel blending.....it seems to be the same thing using a different Ps tool/technique.

And the questions are:
Will channel mixer work OK? and why did you not use it in preference to channel blending via apply image?
How can we use channel mixer to get your channel blending results; i.e. what are the basic procedural steps and perhaps a sample image done both ways?

I'm thinking it will work but I have not been successful doing so, yet. (I have not given up)

Doug Schafer


Thomas Hurd,MD
 

One of my favorite exercises was the Panama image. I took the moire as  a personal challenge to be eliminated.

Even though I completely wiped out the nose on Dan’s face (and it wasn’t the only time my correction was an outlier), the task at hand forced me to find out how to eliminate the problem. 

Finding the suppressor add on was a challenge to track down, and I actually at one point thought about drawing some of the detail back in. It was a little embarrassing to be the only image with an absolutely smooth face.

But in the end, Dan, in his critique, said my colors were fine and said to blend the nose back in with the blue channel IIRC. So a win for the color part, and a win for learning how to correct the face contrast the next time I get that. In fact, it gives me the ability to now take any detail away and just blend another channel for contrast!

Tom Hurd

On Aug 14, 2020, at 7:18 PM, Thomas Hurd,MD via groups.io <tomhurd@...> wrote:

I will say that I have tried channel mixer, but image>apply image is first of all much easier to begin with. 

I have been using Channel Power Tools, which lends itself beautifully to apply image. 

I think that working on the channels for blending and masking is one of the most wonderful thing about Photoshop in image correction.

When I had no idea what I was doing in Carnival, having never produced a CMYK image before, I took the RGB channels and essentially copied them into CMYK. From there I used adjustments like curves, etc. but also the brush tools, like dodge and burn, and the brush itself to paint in shadows and highlights for the individual channels. I Emma’s able to get a better shape to the colored image that way.

I didn’t use the brush techniques on the channels frequently in the other exercises, except during masking.

Tom Hurd

On Aug 14, 2020, at 2:56 PM, k_d@... wrote:

Our thanks again for everything and helpful final words.

And questions:
I think most have learned how to channel blend via apply image to channels and use for image correction....it is a valuable tool as you pointed out in your summary.
I'm thinking we can achieve the same result using channel mixer adjustment layer in place of apply image channel blending.....it seems to be the same thing using a different Ps tool/technique.

And the questions are:
Will channel mixer work OK? and why did you not use it in preference to channel blending via apply image?
How can we use channel mixer to get your channel blending results; i.e. what are the basic procedural steps and perhaps a sample image done both ways?

I'm thinking it will work but I have not been successful doing so, yet. (I have not given up)

Doug Schafer