Topics

Colosseum: Results


Dan Margulis
 

I’ve posted the results of our Cinque Terre study as described below.
 
Reviewing: This is from the MIT study of 5,000 images. It is a night shot of the Colosseum in Rome, which according to Wikipedia, in 2018 was the most visited tourist attraction in the world. Some other information: recently Rome has replaced most of the outdoor lighting with LED technology, which is extremely controversial with the locals, who say it makes the whole area look like the inside of a refrigerator. This shot, however, dates from before the replacement, so the lighting is warm.

I have never used this in a class so I don’t know how difficult experienced people find this. My notes to the study, however, indicate that the MIT retouchers did a good job with it; their average result is among the entries, as is an excellent version by one of the retouchers.

We have a record 29 entrants. Another that came in half an hour after the deadline is not included.  When a person submitted two or versions, I chose the one I thought was better. Most people also submitted a list of their steps, thanks very much. I haven’t read these, because I’d rather get a sense of who was successful and who wasn’t before investigating why. 
 
The files don’t have people’s names on them, and were random-generator numbered from #601 to #629. As with past studies, we also have a “par” version, #630. To get it, I chose what I thought looked like the five best entrants, and averaged them, each one weighted 20%. This often creates a version that is superior to most if not all of its parents.
 
I’m going to have some things to say about this assortment, but as usual I’d like to open it up to group discussion first. What do the successful versions have in common? Meanwhile, if you’d like to know how your own version stacked up, download the par version and compare the two directly. Do you think you got the same kind of quality? If not, I hope you’ll find further discussion useful.
 
The folder is in the group Photos section, named Case Study: Colosseum 
Because some of us would like a closer look at these, I also have zipped all 30 and uploaded a 70 mb file to our Files section,
 
Search for Colosseum_entries_062220.zip
 I look forward to your comments.

Dan Margulis

P.S. The next case study is announced today, look for a separate post.


Kenneth Harris
 

I find 604 very poetic.  There's a lot good to say about 620, but greatly weakened by the acutance along the top which needs either to be softened or balanced with more punch on the Colosseum.  But I keep coming back to 604.

Ken Harris


Gerald Bakker
 

The main challenge for this image, as I see it, is to make the colosseum stand out against the background and surroundings. This can be accomplished in different ways, each of which can work well.

- In #606, it was done by making the background very dark.
- In #613 and #627, by emphasizing color variations.
- In #617, by giving the colosseum a nice warm glow.

Where depth is lacking (e.g. #608 and #621) the image appears flat.

Another choice to make is, what to do with the blue cast in the background. Some versions, e.g. #626, have completely removed it. Some, like #620, emphasized it. My own version is #619. I toned down the blues but kept some of it.

I also think the lit arches should stand out from the rest of the building. In some, e.g. #611 and #620, this is very strong, as though there is a fire burning inside the colosseum.
Anyhow, I love to click through all the versions. Even though they differ considerably, many are a joy to look at.
 
--
Gerald Bakker
http://geraldbakker.nl


Doug Schafer
 

To all,
my thoughts on Colosseum entries:

Overall, in general, I think most got the contrast well, tho a few looked flat.

Some colors look wrong: too yellow, too orange, too blue, or too little (low sat).

#606 did an excellent job of focusing attention on the square and Colosseum and letting the rest be supportive.

As usual, par image was probably best....but I thought almost all missed a big problem of bright street light in near center which distracts too much from the total image. Light should be fixed (reduced brightness).

Most scenes were nicely dusk early evening city luminance; but several too light and several too dark. Many had great shadows but several made the green trees washed out with foggy yellowish light.

Doug Schafer


Doug Schafer
 

To All,

I thought I'd also make a quick comment on every submitted image...my opinion (take it or leave it, just my observations):
I still think most should have decreased brightness of street light at Colosseum; very distracting to rest of image.
Tho rated like par, some of these seem too saturated, by a little

601    on par with par, nice
602    too yellow, Colosseum not color or luminance differentiated from bkgd
603    too reddish lights, bkgd too light
604    overall very good, could use a little contrast and slightly darker bkgd
605    on par with par, nice
606    nice version to give attention to scene with supportive bkgd
607    too blue, too bright
608    flat brownish grey, bkgd too light
609    very nice but foreground shadows much darker than bkgd, trees black
610    a bit flat & grey, needs more color beyond street lights; still attractive
611    too contrasty foreground, too blue, trees too light; no depth from colors, grainy
612    dark but captures night feel well, green trees too foggy orange
613    overall nice but needs darker foreground shadows & bkgd
614    on par with par, nice
615    on par with par, nice
616    on par with par, nice except seems to need more color vibrance (not sat.), a little
617    on par with par, nice...except slightly too bright
618    on par with par, nice but bkgd needs less blue and slightly darker but yields nice color contrast
619    on par with par, nice
620    on par with par, nice, but foreground oranges too saturated, scene could be a bit darker
621    green, grainy, flat, harsh
622    on par with par, excellent, nice attention to street & Colosseum
623    very similar to 622 with less attention to street & Colosseum
624    on par but seems too bright
625    on par with par, nice
626    nice but trees too green & bright, right foreground too bright
627    too much/many blues flattens scene, trees too light, street lights too yellow
628    could be on par but too little color, street lights washed out, flat needs contrast, dull
629    OK but maybe needs less saturation, little less blue, touch more contrast
630    par excellence

Doug Schafer


Robert Wheeler
 

In looking at the Colosseum image, I found the original to have muted colors and inadequate contrast. I thought it would be interesting  to enhance the warm colors of the lower part of the image and the cool colors of the upper part of the image for artistic color contrast. The PAR version (630) does this.

 

Contrast responded well to my initial efforts, but the colors were not easy fixes, with each adjustment helping some areas but worsening others. Eventually I started over and purposely made a black and white version using NIK Silver Effex Pro 2 and NIK HDR Effex Pro 2 (single image method). The result presents a good amount of sparkle and liveliness, but it seems out of keeping with the intent of a color retouching challenge. Perhaps headed in a similar direction, entry 606 minimized color while maintaining good visibility of the Colosseum wall. It is enjoyable when viewed away from the other entries. Entry 628 was also in the less colorful group, but somewhat darker. I like full B&W a little better.

 

I started over following PPW methods. Converted to LAB, took a slight amount of green out of the sky, ran the shadows and highlights action, selected the yellow-orange areas for “skin” desaturation then applied the MMM-CB action (way too saturated without the earlier “skin” desaturation action), worked on the sky color with increased saturation and decreased lightness of blue and cyan, did final curves auto adjustment in luminosity mode with inverted mask painted to apply the enhancement  to street area, save as jpg. This maintained warm colors in the lower areas and cool colors in the upper areas, provided good detail in the street action,  and is the largest component of my entry.

 

A number of entries kept warm colors in the lower area but had the upper area more dark than cool (606, 610, 612, 616, and 621 to my eye), with some of these omitting or minimizing skyline building shapes or the strings of road lights near the top. Some good images here, but perhaps a missed opportunity artistically. To me the bright yellow foreground of 602 is not very warm feeling and has less color contrast with the  very dark sky.Entry 607 has a very cool background, but the cyan/blue extends over the Colosseum wall making an uncomfortably artificial (or perhaps “artistic” effect) where I would expect the artificial lighting to make for neutral and warm tones instead.

 

My last processing step was to layer my B&W version onto my PPW version at 50% opacity in luminosity mode, which improved the brightness/darkness dynamics, lightened the image overall, and enhanced details. This is 615. In comparison to the others, I see it could benefit from being a little brighter overall and perhaps a little less red/magenta. Blending the PAR at 25% to 50% helps.

 

The remaining 19 entries have varying degrees of cool backgrounds and warm foregrounds and varying degrees of brightness of the Colosseum wall and definition of the street activity. To my eye the following have good strengths: 623, 614, 617, 620, 625, 604, 624, 609, 603, 613, and 615.

 

Robert Wheeler


Dan Margulis
 


On Jun 22, 2020, at 7:53 AM, dmargulis <dmargulis@...> wrote:

I’ve posted the results of our Cinque Terre study as described below.

A new addition to the folder will be of interest to students of par images and averaging generally.

Through all these case studies we consistently have liked the par versions and often considered them the best ones in the folder. As you know, I produce these by quickly selecting what I would guess to be the best five versions (rather tough when we get up to 29 entrants) and then averaging them, with each one given 20% weighting in the par version.

Harvey Nagai has taken the trouble to go further: he has averaged all 29 entrants. I have uploaded this file to the folder as #631.

Toggling back and forth between the “real” par version #630 and this new #631, well, clearly #630 is somewhat better, but you might be surprised by how small the difference is. After all, among these 29 images are some, shall we say, real losers, whereas even if you don’t agree with my specific choices of five for the par version, the comments so far are that they are five good ones.

The answer: in averaging, errors tend to cancel each other out. Among the 29 are some that are much too light but others that are much too dark. Some are too warm and others too cold, etc.

This effect also explains why we like the “real” par versions so much. We see them as free from error. After all, as a group we are very good at finding trivia to complain about. For example, here are three versions that have at least two favorable mentions from commentators, and that I also feel are excellent. According to me, they also have the following flaws:
#606 in trying for a very dark background went slightly overboard and plugged stuff that shouldn’t be plugged.
#617 in trying for a nice, light golden look in the Colosseum, blew out its lightest areas.
#620 has a nice break between foreground and background, but the Colosseum is too neutral.

If you average these three, the above three defects go away, or at least become invisible. What *is* visible is what they have in common, namely, excellent separation of the Colosseum from the background city.

Thanks to Harvey for this demonstration.

Dan



Jan Hof
 

On Mon, Jun 22, 2020 at 08:33 PM, <k_d@...> wrote:
As usual, par image was probably best
Yep, after reviewing all the edits, this is also my conclusion. The background sky has a nice dark blue. And the moment I saw that blue, I realized that I did want such a blue background, but did not accomplish that!
Other favorites are #601, #606 (the plugged trees didn't bother me, until Dan mentioned it; and then it is not possible to not see the trees as plugged anymore) and #619

My edit (#622) has a road with streetlights that are too red; after seen other edit, I would have changed the color more to yellow. But apparently, when you edit long enough it is hard for me to keep a fresh attitude and not get used to colors.

------
Jan Hof


David
 

Hi,
I perpetrated #606. Thank you for the comments. I do appreciate the feedback.

Actually, when I processed it originally, the blacks weren't plugged.

I stopped using Photoshop, and later Lightroom, after Adobe forced everyone into the rental model. Maybe two years ago, I finally settled on ON1 as a digital darkroom tool. So when I decided to take a shot at the Colosseum, I didn't know whether a non-Photoshopped entry would be acceptable.

I processed the picture taking as marching orders that the scene should be for Rome's tourist board. That led me to focus on the arena and the lit pedestrian area (including the trees). To get the processed image out to Dan, I had to export my work as a jpeg. I noticed that the exported image had taken dim shadows and converted them to solid black. Since I wasn't sure that an ON1 entry would be an acceptable entry, I didn't go back to lighten the black areas. Mentioned to Dan when I sent the image that if it was ok to use non-Photoshop product to let me know and I would fix the blacks.

So, I'll know for next time, that I can get away with using ON1 and only to submit what I'm satisfied with.

Incidentally, if anyone is interested, I think that ON1, which can also work as a Photoshop plug-in, is a good product. It's not as polished as Adobe's products, but it's getting noticeably better with each release.

Again, thanks for the comments and thank you Dan for accepting my entry.
David


Dan Margulis
 



On Jun 22, 2020, at 2:05 PM, Gerald Bakker <gc.bakker@...> wrote:

The main challenge for this image, as I see it, is to make the colosseum stand out against the background and surroundings. This can be accomplished in different ways, each of which can work well.

- In #606, it was done by making the background very dark. 
- In #613 and #627, by emphasizing color variations.
- In #617, by giving the colosseum a nice warm glow.

This is an excellent analysis. We have to juggle two priorities. First, the subject is one of the world’s great attractions, and we must emphasize it. Second, the background can distract from it, yet we can’t just make it go away. Rome at the time the Colosseum was built ruled the civilized world. Today, it is a vibrant city of three million. Our rendition should reflect this.

The majority (16 of us, by my count) adopted a strategy of having some kind of warm look in the foreground, including bright lights inside the Colosseum, set off by a significantly darker background, often blue. I’m going to ignore the other versions in this post and concentrate on what makes this strategy work.

First, three minor improvements that most didn’t try.
1) Some kind of spotlight is shining on the right side of the Colosseum, affecting the second and third level of arches. Several people emphasized that lighting in one way or another, making the whole scene more dramatic.

2) Understanding that darkening the background reduces its detail, a few people tried to compensate by adding color to the buildings at left rear. This was done particularly effectively in #609 and #627.

3) The choice of blue as a background color is probably better than a neutral, but it does give rather a cool overall look. #601 decided that a more purple sky would set the mood better and I think he was right.

Now, on to the nitty-gritty. PPW is not especially useful in this particular exercise. Oh, you can use it, but it doesn’t have great advantages over other methods. We saw more of that in Cinque Terre, and will se even more in the next MIT exercise. So we’re back to the basics. Here is a listing of the 16 versions that, according to me, are trying to implement the strategy mentioned above.

#601 (72,16) 56
#603 (66,33) 33
#604 (52,28) 24
#606 (65,16) 59
#609 (54,23) 31
#610 (58,24) 34
#612 (54,13) 41
#614 (59,18) 41
#615 (47,17) 30
#617 (79,19) 60
#619 (63,24) 39
#622 (54,18) 36
#623 (52,22) 30
#625 (53,29) 34
#626 (61,10) 51
#629 (56,21) 35

The numbers accompanying each are L values—the smaller, the darker. The first represents the lightest part of the Colosseum other than what’s in the spotlight. I find it in the center of the Colosseum, right above the top level of arches. Note that this is not a highlight, which would be found somewhere in the area in front of the structure.

The second number is a section of the sky above and slightly to the right of the first point. Note that it is not a shadow; many other areas of the background are darker.

And the third is the difference between the two. Let’s now have a look at the list in that order:

#617 (79,19) 60
#606 (65,16) 59
#601 (72,16) 56
#626 (61,10) 51
#612 (54,13) 41
#614 (59,18) 41
#619 (63,24) 39
#622 (54,18) 36
#629 (56,21) 35
#610 (58,24) 34
#625 (53,29) 34
#603 (66,33) 33
#609 (54,23) 31
#615 (47,17) 30
#623 (52,22) 30
#604 (52,28) 24

If I’ve never seen any of these versions I’d be willing to bet that the top four are the four “best”. In fact, my own favorites are #617 and #601. I downgrade #606 for crushing the left half of the image and #626 for being too orange, but I’d take either one in a heartbeat over anything beneath them on the list.

I suggest that if you take any image on the above list and compare it to another where the difference is >5, you’ll prefer the one with the greater difference.

If you doubt it, try fixing up yours or one of the others on this list. It seems to me that the first number should be 65 or higher and the second 20 or lower. If neither one is so, a curve like the one attached is appropriate. If the first number is acceptable but the second too light, then eliminate the point that lightens the midtone. And if only the first number is too low then just write a curve to lighten the midtone. This is NOT the way I advocate doing things, but it should substantially improve any image not meeting the above numerical suggestions.

Dan Margulis

P.S. Note that a curve like this is normally not acceptable because it completely flattens the shadows. Here, there isn’t any shadow detail to protect, so it should be fine.






ROBIN MARK D'ROZARIO
 

Dear Dan,

Thanks for the comments and analysis.
I wish I could claim intent for the luminosity range of my submission, 617. It was more happenstance and luck.
The submission is a blend of three images. One of these was used only for the street and trees so it didn't contribute to the Colosseum and background.
Of the other two one has a range of over 72 points and the second around only 41! Without the contribution of the former my submission would have been quite flat. I actually thought that the second image was more balanced and a better choice by itself! A further endorsement for blending images.

A couple of points of enquiry:
1) This image is without any obvious neutral shadow and highlight areas with detail. How do you manage colour here? I made some arbitrary choices here for the colour curve as well as setting Black / White / Grey points, which seem to have worked out fine, but I might have just got a lucky break.
2) What is the rationale behind the choice of the two points you used to highlight Luminosity range?

Best regards,
Robin Mark D'Rozario


Thomas Hurd,MD
 

Dan,

Thank your the analysis, and reminding us that good contrast and believable color count most. I was shocked to find that you published a curve with dark to light going from left to right, instead of right to left!

I found it interesting that almost all the variation from the default to the par occurs in the upper half of the image. 

Realizing this, I just now tried to recreate par by
  • 1 taking the default into camera raw and applied the graduated filter with dehaze +62 and exposure =0.35.
  • 2 I added a CRV increasing luminence of the colosseum face, and 
  • 3 a brightness/contrast layer to darken the background through a red channel mask, inverted. 
  • 4 That got me about 80% to par. 
  • 5 Some additional adjustment layers, esp hue saturation, and
  • 6 a luminosity layer with black channel applied to RGB got even closer.

Since your comments on the last exercise, I have made a point to try to recreate the par images from scratch, and find it helpful. I am one of the less experienced that travels down some unfruitful paths, so trying it again afterward is even more instructive.
  
In my own entry, 626, I tried to separate 5 sections of the picture: 
  • 1 colosseum, 
  • 2 background with town and sky, 
  • 3 trees, 
  • 4 street and
  • 5 foreground column. 
My intent was:
  • 1 colosseum differentiated from background and trees
  • 2 bring green into the trees
  • 3 Try to approach neutrality in the street between light reflections 
  • 4 keep the foreground column recognizable and not plugged
  • 5 darken the background.
My methods:

1 Use individual curve layers and find a channel to modify and mask out the areas of interest/disinterest with heavily blurred masks
2 Putting those masked puzzle pieces together constituted my color and contrast steps for the image.
  • I agree it was nearly impossible to just pick a highlight and shadow, and proper colors (that the viewer would accept) were somewhat arbitrary, but they had to reflect the time of day and artificial lighting. 
  • As I went through the process, I had several intermediate steps with more blue in the sky and town, but preferred the look after the blue hour, just after astronomical dusk, approaching black.
3 After that I followed a more regular PPW approach, using H-K, Bigger hammer and then MMM/CB with the top floor of the colosseum as my loosely lassoed selection.
4 I used some different hue/saturation and curves adjustment layers to put the final touches on.
5 I did use two spotlight methods for the wall. 
  • First just a black fill layer with white circular brush dabs in the layer’s inverted mask.
  • Second, I used a radial gradient fill (diamond) with orange in the center and blue to the outside.
6 Time: I spent 3 plus hours on Sunday in-between home chores to get the image. Over two days I saved about 13 individual images, and on Sunday 6 images in the step by step by step process. A good hour of my time was spent trying to reconstruct my steps.

Results:
  • I achieved my primary goals for the image. I agree the center is more orange than other entries, as you pointed out, and the trees are a little bright green as another noted. 
  • I was happy with the overall contrast and weight but would have liked to get more warm color in the wall, rather than the street. Although I liked the green in the trees, I could not manage enough green variation to give the impression of depth, and it became a little more blotchy than I wanted.
  • After the submission I went back and tried some other methods to give more definition to the trees, bur none were quite satisfactory.
Everybody else:
  • I really enjoyed everyone’s interpretations. 617 was my personal favorite for the warm glow noted by Gerald
  • Like the last exercise, analyzing each image highly depended upon the previous image! 
  • I have to say 607 reminds me of “CAFE TERRACE AT NIGHT” 

Tom Hurd

On Jun 23, 2020, at 10:56 AM, Dan Margulis via groups.io <dmargulis@...> wrote:



On Jun 22, 2020, at 2:05 PM, Gerald Bakker <gc.bakker@...> wrote:

The main challenge for this image, as I see it, is to make the colosseum stand out against the background and surroundings. This can be accomplished in different ways, each of which can work well.

- In #606, it was done by making the background very dark. 
- In #613 and #627, by emphasizing color variations.
- In #617, by giving the colosseum a nice warm glow.

This is an excellent analysis. We have to juggle two priorities. First, the subject is one of the world’s great attractions, and we must emphasize it. Second, the background can distract from it, yet we can’t just make it go away. Rome at the time the Colosseum was built ruled the civilized world. Today, it is a vibrant city of three million. Our rendition should reflect this.

The majority (16 of us, by my count) adopted a strategy of having some kind of warm look in the foreground, including bright lights inside the Colosseum, set off by a significantly darker background, often blue. I’m going to ignore the other versions in this post and concentrate on what makes this strategy work.

First, three minor improvements that most didn’t try.
1) Some kind of spotlight is shining on the right side of the Colosseum, affecting the second and third level of arches. Several people emphasized that lighting in one way or another, making the whole scene more dramatic.

2) Understanding that darkening the background reduces its detail, a few people tried to compensate by adding color to the buildings at left rear. This was done particularly effectively in #609 and #627.

3) The choice of blue as a background color is probably better than a neutral, but it does give rather a cool overall look. #601 decided that a more purple sky would set the mood better and I think he was right.

Now, on to the nitty-gritty. PPW is not especially useful in this particular exercise. Oh, you can use it, but it doesn’t have great advantages over other methods. We saw more of that in Cinque Terre, and will se even more in the next MIT exercise. So we’re back to the basics. Here is a listing of the 16 versions that, according to me, are trying to implement the strategy mentioned above.

#601 (72,16) 56
#603 (66,33) 33
#604 (52,28) 24
#606 (65,16) 59
#609 (54,23) 31
#610 (58,24) 34
#612 (54,13) 41
#614 (59,18) 41
#615 (47,17) 30
#617 (79,19) 60
#619 (63,24) 39
#622 (54,18) 36
#623 (52,22) 30
#625 (53,29) 34
#626 (61,10) 51
#629 (56,21) 35

The numbers accompanying each are L values—the smaller, the darker. The first represents the lightest part of the Colosseum other than what’s in the spotlight. I find it in the center of the Colosseum, right above the top level of arches. Note that this is not a highlight, which would be found somewhere in the area in front of the structure.

The second number is a section of the sky above and slightly to the right of the first point. Note that it is not a shadow; many other areas of the background are darker.

And the third is the difference between the two. Let’s now have a look at the list in that order:

#617 (79,19) 60
#606 (65,16) 59
#601 (72,16) 56
#626 (61,10) 51
#612 (54,13) 41
#614 (59,18) 41
#619 (63,24) 39
#622 (54,18) 36
#629 (56,21) 35
#610 (58,24) 34
#625 (53,29) 34
#603 (66,33) 33
#609 (54,23) 31
#615 (47,17) 30
#623 (52,22) 30
#604 (52,28) 24

If I’ve never seen any of these versions I’d be willing to bet that the top four are the four “best”. In fact, my own favorites are #617 and #601. I downgrade #606 for crushing the left half of the image and #626 for being too orange, but I’d take either one in a heartbeat over anything beneath them on the list.

I suggest that if you take any image on the above list and compare it to another where the difference is >5, you’ll prefer the one with the greater difference.

If you doubt it, try fixing up yours or one of the others on this list. It seems to me that the first number should be 65 or higher and the second 20 or lower. If neither one is so, a curve like the one attached is appropriate. If the first number is acceptable but the second too light, then eliminate the point that lightens the midtone. And if only the first number is too low then just write a curve to lighten the midtone. This is NOT the way I advocate doing things, but it should substantially improve any image not meeting the above numerical suggestions.

Dan Margulis

P.S. Note that a curve like this is normally not acceptable because it completely flattens the shadows. Here, there isn’t any shadow detail to protect, so it should be fine.

<Curve.jpg>






Dan Margulis
 



On Jun 23, 2020, at 2:15 PM, ROBIN MARK D'ROZARIO <rdrozario@...> wrote:

A couple of points of enquiry:
1) This image is without any obvious neutral shadow and highlight areas with detail. How do you manage colour here? I made some arbitrary choices here for the colour curve as well as setting Black / White / Grey points, which seem to have worked out fine, but I might have just got a lucky break.

It’s quite difficult in this original, with garish lighting in the foreground and a blue cast in the background. The Colosseum is known to be somewhat warm and so is the street lighting. Can find potential shadow points without difficulty but there is inevitably some guessork.

2) What is the rationale behind the choice of the two points you used to highlight Luminosity range?

The lighter point is the lightest area of the building other than the spotlighted part, which we avoid for the same reason we avoid using catchlights as our highlight points.

The darker point is something representing the darker area of the sky yet not so dark as a full shadow. Forcing it darker creates the impression of a darker sky. To see the importance of the two points, compare #612 and #614, which are tied for fifth place for the most contrast, well behind the top four. Their treatment is similar. To derive #614 from #612 we could just apply a curve lightening the midtone, but this lightens the sky as well. Therefore as presented I have no preference between them. Buf if instead the lightening curve had placed a holding point at the chosen dark point then #614 would be preferable. And in reverse, if one applies a curve to #612 that lightens the lighter point but holds the dark sky where it is, that version would be preferable too.

Dan



Paco
 

Hi! Mine is 601

 

I had been looking for an excuse to try and process an image as much as I could using only Capture One. I processed this image using Capture One's 16 allowed layers for all corrections. When I ran out of the permitted 16 layers, I applied one last curve and some color to the street in PS.

 

I really like C1’s masking by color range and refining masks is easy. Masks show up as a red overlay. It is easy then to work with them brushing in or erasing what to show, hide or make transparent.

 

I approached this image trying to imagine it as a painting.

 

Don’t know if you can notice that I purposely darkened or eliminated a lot of small highlights such as the 2 at the home on the far right.

 

I still shoot dawn and dusk shots and the sky color which is liked the most is a violet bluer than magenta.

 

What took me the longest was deciding how much to show of the trees.

 

All the best!


John Furnes
 

I found this simple analysis quite remarkable. I tried it on my own version (#623), and though I didn’t get the colours right, I could clearly see how the image was improved ,namely there came more distance between dark and light and a better definition of the different part of the image.

 

I strive with having my images a bit flat, and even though I try to boost colours, it just won’t snap in.

I think I over-used the H-K on the background – to ‘hide’ it, and thereby lost the blue. I realise that magenta is one of my problems, and am working on it.

 

Also, the idea of making several versions, and then combine them appeals to me, as I often do make several versions. I know Dan has said this several times, in his books and here, and I HAVE tried it, but not on a regular basis. I will start trying this approach.

 

Best

 

John Furnes

 

From: colortheory@groups.io <colortheory@groups.io> On Behalf Of Dan Margulis via groups.io
Sent: 23. juni 2020 16:56
To: colortheory@groups.io
Subject: Re: [colortheory] Colosseum: Results

 

 



On Jun 22, 2020, at 2:05 PM, Gerald Bakker <gc.bakker@...> wrote:

 

The main challenge for this image, as I see it, is to make the colosseum stand out against the background and surroundings. This can be accomplished in different ways, each of which can work well.

- In #606, it was done by making the background very dark. 
- In #613 and #627, by emphasizing color variations.
- In #617, by giving the colosseum a nice warm glow.

 

This is an excellent analysis. We have to juggle two priorities. First, the subject is one of the world’s great attractions, and we must emphasize it. Second, the background can distract from it, yet we can’t just make it go away. Rome at the time the Colosseum was built ruled the civilized world. Today, it is a vibrant city of three million. Our rendition should reflect this.

 

The majority (16 of us, by my count) adopted a strategy of having some kind of warm look in the foreground, including bright lights inside the Colosseum, set off by a significantly darker background, often blue. I’m going to ignore the other versions in this post and concentrate on what makes this strategy work.

 

First, three minor improvements that most didn’t try.

1) Some kind of spotlight is shining on the right side of the Colosseum, affecting the second and third level of arches. Several people emphasized that lighting in one way or another, making the whole scene more dramatic.

 

2) Understanding that darkening the background reduces its detail, a few people tried to compensate by adding color to the buildings at left rear. This was done particularly effectively in #609 and #627.

 

3) The choice of blue as a background color is probably better than a neutral, but it does give rather a cool overall look. #601 decided that a more purple sky would set the mood better and I think he was right.

 

Now, on to the nitty-gritty. PPW is not especially useful in this particular exercise. Oh, you can use it, but it doesn’t have great advantages over other methods. We saw more of that in Cinque Terre, and will se even more in the next MIT exercise. So we’re back to the basics. Here is a listing of the 16 versions that, according to me, are trying to implement the strategy mentioned above.

 

#601 (72,16) 56

#603 (66,33) 33

#604 (52,28) 24

#606 (65,16) 59

#609 (54,23) 31

#610 (58,24) 34

#612 (54,13) 41

#614 (59,18) 41

#615 (47,17) 30

#617 (79,19) 60

#619 (63,24) 39

#622 (54,18) 36

#623 (52,22) 30

#625 (53,29) 34

#626 (61,10) 51

#629 (56,21) 35

 

The numbers accompanying each are L values—the smaller, the darker. The first represents the lightest part of the Colosseum other than what’s in the spotlight. I find it in the center of the Colosseum, right above the top level of arches. Note that this is not a highlight, which would be found somewhere in the area in front of the structure.

 

The second number is a section of the sky above and slightly to the right of the first point. Note that it is not a shadow; many other areas of the background are darker.

 

And the third is the difference between the two. Let’s now have a look at the list in that order:

 

#617 (79,19) 60

#606 (65,16) 59

#601 (72,16) 56

#626 (61,10) 51

#612 (54,13) 41

#614 (59,18) 41

#619 (63,24) 39

#622 (54,18) 36

#629 (56,21) 35

#610 (58,24) 34

#625 (53,29) 34

#603 (66,33) 33

#609 (54,23) 31

#615 (47,17) 30

#623 (52,22) 30

#604 (52,28) 24

 

If I’ve never seen any of these versions I’d be willing to bet that the top four are the four “best”. In fact, my own favorites are #617 and #601. I downgrade #606 for crushing the left half of the image and #626 for being too orange, but I’d take either one in a heartbeat over anything beneath them on the list.

 

I suggest that if you take any image on the above list and compare it to another where the difference is >5, you’ll prefer the one with the greater difference.

 

If you doubt it, try fixing up yours or one of the others on this list. It seems to me that the first number should be 65 or higher and the second 20 or lower. If neither one is so, a curve like the one attached is appropriate. If the first number is acceptable but the second too light, then eliminate the point that lightens the midtone. And if only the first number is too low then just write a curve to lighten the midtone. This is NOT the way I advocate doing things, but it should substantially improve any image not meeting the above numerical suggestions.

 

Dan Margulis

 

P.S. Note that a curve like this is normally not acceptable because it completely flattens the shadows. Here, there isn’t any shadow detail to protect, so it should be fine.