Did this scan come out alright?

[William Michael]

Which is better, negative or print? The negative, if you have it. . . A print has about 3.5 stops dynamic range, or less if old and faded or poorly exposed. . .

 

I agree with those who believe this was scan of a print; but I still would like the OP to confirm

 

When working with an old print for restoration, we would often retouch the print first before making an inter-negative; scan or digital (camera) copy file. Those who have those retouching skills are probably now few and far between,much less having the sets of dye inks and air brushes for the retouching and/or rebuilding of a faded print. The cost to do that work is a consideration also, but in many ways it is easier to rebuild the original print by retouching/airbrushing, for it to have more 'guts' with which to work later digitally on the scan or copy file. That's the main reason why I mentioned that 'better' results could be attained by working from the original.

 

I am not convinced that the scan (of the print) is the best possible - perhaps the OP might disclose the details of that scanning process - there might be suggestions how that could be improved.

 

WW

[dcstep]

...

 

The image I adjusted using Photoshop Levels is the 8th post in this thread. Below is adjusted using the white dropper in the upper border, which is evenly faded but not stained.

...

 

If you'd try the dropper, where it belongs, on something close to medium grey, on the subjects, like the shadows on the woman's blouse, you'd get a more pleasing rendition. Why don't you do it, just to prove me wrong?

[Tim_Lookingbill]

Here's my take on color correction. The image is too bright so I applied a custom point curve in ACR 6.7. Adjusted white balance and stained the areas on the walls where there's a lot of magenta by applying an adjustment brush set to negative saturation and red orange tint. Lots of HSL to get skin color right.

 

[Gary Holliday]

Interesting to see everyone's interpretation of the print. Some of the interpretations lean towards magenta, especially the lady on the right. I personally prefer the warmth and age of the original silver halide print, the digital versions can look harsh and technical, but dcstep's is a nice balance of the two with good blacks.

 

Anyone notice the cheeky 'over the shoulder' of the professional photographer shot :) Naughty naughty.

[Ed_Ingold]

If you'd try the dropper, where it belongs, on something close to medium grey, on the subjects, like the shadows on the woman's blouse, you'd get a more pleasing rendition. Why don't you do it, just to prove me wrong?

There is no definitive answer. Photos of this sort never looked real, or even good compared to what we expect today from digital.

 

There is no neutral for grey balance. Even if there were, all you achieve is red/blue balance. Shadows are colored by the surroundings and the surface on which they fall. Sampling the border removes most of the staining due to age. Everything else is putting lipstick on the pig.

 

This doesn't mean there isn't sentimental value in a photo taken (my guess) 45-50 years ago. Rembrandt (or Wesson), it isn't and never will be. Please feel free to dazzle us with your grading proficiency.

Edited July 11, 2017 by Ed_Ingold

[dcstep]

Ed, when I look at the unadjusted image, I see a very large difference between the border and whites of the shirt and blouse. I think that we're dealing with a scan of a print here, so when it was printed, that border would have received no light (data) of any kind. The image within the border received light, adjusted by the negative, which caused a different chemical reaction than the border, which received none. When new, the border would have looked the same as an unused piece of printing paper.

 

Decades later, some of us are trying to achieve a "modern" interpretation of the colors, without the built in color flaws caused by the lighting, compromises in the negative stock and compromises in the paper stock. Even with perfect WB, back in the day, negative and positive films each had their own "character", which tended to vary quite a bit from what our eyes saw. Anyway, modern technology does allow us to get closer to what our eyes would have seen, had we been there. The first step, with most software, is the eye dropper placed on a neutral grey portion of the subject. The border was not influenced by the lighting or the negative stock, but the subject was and that's what we're correcting.

 

Yes, it's subjective. We start with the dropper. In this case, I felt no need for further adjusting, but others might. With my own work, I often adjust what the dropper predicts, but I still start with the dropper on a neutral grey.

 

Up the thread, someone mentioned that scanning technique might have been better. I agree, mainly because it seems OOF, as if the print weren't totally flat on the scanner. A pro photographer might have made an OOF shot, but I'm doubting it, in those circumstances. He was far enough away that f/8 would have focused that whole scene and I'd bet that he or she used an even smaller aperture with the lights of the day. I think the OP should try to rescan, paying particular attention to getting the print flat on the scanner. I guess that it will sharpen up substantially.

[Tim_Lookingbill]

Interesting to see everyone's interpretation of the print. Some of the interpretations lean towards magenta, especially the lady on the right. I personally prefer the warmth and age of the original silver halide print, the digital versions can look harsh and technical, but dcstep's is a nice balance of the two with good blacks.

 

Anyone notice the cheeky 'over the shoulder' of the professional photographer shot :) Naughty naughty.

 

OK, looks too magenta? Here's another edit...

 

[Tim_Lookingbill]

There is no eyedropper neutral setting to correct for color balance on this image. There are so many color constancy distortions due to the harsh magenta/blue arc lamp color temp type lighting this scene was lit by that screws up hues in neutrals and primary memory colors.

 

This could be caused and/or exacerbated by the non-linear over brightening of the scan that I've seen brightening my own DSLR Raw files where any color tint combo nearing clipping in the highlights throws off color balance to produce a kind of staining, in this image it's on the magenta side.

Edited July 12, 2017 by Tim_Lookingbill

[dcstep]

Tim, I think that your last color interpretation is the best so far.

[Bill C]

Most of the posted corrections look sort of ok, but I think this is only because they are being seen alone - no side by side comparisons. People can adapt pretty drastically when they have only a single image to look at, especially on a monitor.

 

In reality there are some significant color problems with all of the fixes so far - they are most apparent with the woman on the right, who started out with a fairly pale complexion. In the corrections, she has mostly had the color sucked out of her, giving her skin a ghastly cold-blue tone. This would be more obvious on a physical print, which is hard to isolate from its surroundings.

 

To make this more apparent here, I adjusted from the original scan, clipped out the right-side woman, and displayed it side-by-side with another example from post #24 (hopefully I can get this to work).

Ps, I haven't looked at Tim's latest, it's not included in my judgment.

Edited July 12, 2017 by Bill C

[Tim_Lookingbill]

A lot of that pale skin on the woman on the far right is due to her white dress under her chin reflecting back up to her face the harsh blue violet tinted flash. You have to warm everything up to override it or adjust HSL for skin. It's quite difficult to see yellow/orange tint in skin with bluish light. There's also tan lines to consider as well which both women seem to show.

[Bill C]

A lot of that pale skin on the woman on the far right is due to her white dress under her chin reflecting back up to her face the harsh blue violet tinted flash.

 

No, I disagree (except for the part about the flash being harsh). More light being reflected up doesn't cause skin to become pale (although if printed much too light this would be the appearance). It would be hard to convince me of problems with white shirts as I have largely made a living in high volume portrait work without it being an issue. I would guess that she simply has a lighter complexion (possibly the other two are more tanned), and is thus more affected by the print fading.

 

Visually, yellow dye fading is somewhat counteracted by the build-up of yellow "stain." Then if someone tries to "zero out" the yellowish paper base, the loss of yellow dye is more obvious. (Many years ago, when the industry largely switched from white-bordered prints to borderless, it was seen as a secret advantage that customers would lose that white reference, making them less critical about color.)

[Tim_Lookingbill]

I'm not talking about yellow stain, Bill.

 

That woman's face when I was adjusting WB and HSL magenta slider showed non-uniform facial skin while the other woman didn't. This could be a combo of the light used, her position closer to the light with the white top causing her to get the brunt of most of the flash as a hot spot and along with her complexion. Also, once neutralized to some extent reveals quite a few magenta tinted stains in the neutral walls and floor next to the red rug that the yellowing concealed. Either the print processing was subpar back then or the stains are from non-uniform aging of dyes.

 

I've restored some of the most crappy prints made in the '60's that never had so many wonky color stains as this one. It's why I had to use an adjustment brush and edit locally for just this image because color constancy was wack.

[Bill C]

Also, once neutralized to some extent reveals quite a few magenta tinted stains in the neutral walls and floor next to the red rug that the yellowing concealed. Either the print processing was subpar back then or the stains are from non-uniform aging of dyes.

 

Hi, I think that the high base stain levels are somewhat uneven; although I haven't posted it, my full-size adjusted image has a lot of variability in the "white" borders - much more than is apparent in the original scene. (I presume it is more obvious due to "amplification" of weak color "signals.")

 

Note that when I say "stain," I'm using the photo industry meaning - the "white" base color, or density, not the colored streaks due to something having been spilled, etc.

 

I think that this stain level has a lot to do with the uneven colors for two reasons - first is that the yellowish stain is considered to come largely from unreacted magenta dye couplers in the paper, and secondly, that the oxidation of these couplers was probably not consistent due to print storage (being stacked, etc.). The significance of the first part is, that where magenta dyes exists, we expect less yellowish stain (there is less unreacted my coupler).

 

For example, in my adjusted image (post #35), the right side woman's arm has a strong yellowish to reddish (or magenta) shift. This is possibly explained by less underlying yellow stain in the reddish skin. (This is just a guess, not confirmed.) I didn't do local corrections, only pushing and pulling on "curves," as gently as I could.

[Bill C]

A lot of that pale skin on the woman on the far right is due to her white dress under her chin reflecting back up to her face the harsh blue violet tinted flash.

 

Regarding the "blue violet tinted flash," here is a comparison of the right side woman from your post #32 vs my adjustment. On mine, she didn't pick up bluish highlights in her hair, etc., so I think you have to look at how image processing could have put it there.

 

[Tim_Lookingbill]

Regarding the "blue violet tinted flash," here is a comparison of the right side woman from your post #32 vs my adjustment. On mine, she didn't pick up bluish highlights in her hair, etc., so I think you have to look at how image processing could have put it there.

 

I can assure you post processing did not put the blue violet tint in the woman's hair. It's there in the image due to the flash. This is typical of almost every flash exposed dark haired subject in a print I've experienced since I shot film starting in the '70's and on up. Your editing on the right sacrifices the memory color of the seafoam colored dress turning to urine yellow in order to neutralize the flash caused iridescent reflection in the hair.

 

I gave it another go and noticed I surprisingly applied way too much green tint WB and blue so I added two clicks each color temp/tint sliders to add orange but it didn't get rid of the iridescent shine in the woman's hair. As I've said before the color constancy is so wack on this image it takes several go rounds just to fight adaptation.

[Bill C]

Your editing on the right sacrifices the memory color of the seafoam colored dress turning to urine yellow ...

 

You don't think the original scan color of the dress is right? We'll never know for sure, I guess.

 

Here's a 3-way comparison, original scan, my adjustments, and yours. What I did, basically, was to remove some yellow from the (presumably) white clothing, and "blacken" the (presumably) black clothing. Then I looked at the flesh tones, collectively, then pushed and pulled the individual RGB "curves" to move the skin tones towards a more typical set of values (in sRGB). I did not do any hue rotations nor shifts, nor any localized tweaks - only global changes to the RGB curves.

 

My adjustments have exaggerated, in the skin tones, the yellowish to reddish/magenta transitions (visible in a close look at the man's face - yellowish between the brows, then getting reddish immediately above the yellow zone.). Largely it's just an exaggeration of slight color differences. From my experience, these sRGB flesh tone values would print nicely on a well-behaved printer. I wouldn't call it pro-level work, but I don't think many people would complain about it.

 

I've worked with this sort of thing for a living, on a very large scale. I can summarize a little experience if you'd like, but loosely I was working full-time as a photographer about 1970-71, and never left the business (until a couple years ago, maybe temporary, maybe not.) The majority of my experience is technical/lab work, often dealing with color problems.

 

Back to the subject at hand, I've been showing clipped bits, so here's the location of the original, currently in my pnet portfolio: (it would be a lot more palatable with the discolored borders clipped off; they leave a bad impression on the viewer, even though the color within the image is fairly decent)

 

Edited July 13, 2017 by Bill C

[Tim_Lookingbill]

My adjustments have exaggerated, in the skin tones, the yellowish to reddish/magenta transitions (visible in a close look at the man's face - yellowish between the brows, then getting reddish immediately above the yellow zone.). Largely it's just an exaggeration of slight color differences. From my experience, these sRGB flesh tone values would print nicely on a well-behaved printer. I wouldn't call it pro-level work, but I don't think many people would complain about it.

 

If you're going to this much trouble to only go what I'ld say 1/4 of the way to a full restoration, I don't see the point in even making a print with this much yellow of a scene that's clearly been lit by flash made obvious in the shadows and not tungsten.

 

I also don't know the overall point in you're last post.

 

But to answer the OP's topic, I'ld have to say no, this scan did not come out right.

Edited July 13, 2017 by Tim_Lookingbill

[Kent Shafer]

removed flash shadows

 

Removing those flash shadows made a big difference. Could you please explain how you did it?

[Wilmarco Imaging]

Great results to all and very educational for those of us, such as myself, who can use some postprocessing skills improvement.

[William Michael]

Removing those flash shadows made a big difference. Could you please explain how you did it?

 

Mainly, I used the "Clone" tool in Photoshop.

 

I have been told that there are other methods to achieve a similar outcome, but I learned how to use many aspects of the Clone Tool from the outset of my conversion to Digital Photography and thus, the Digital Darkroom processes. Photoshop was my first digital editor and I think that I often default to it because of that, often simulating a mimic of how I would have used film enlarging/printing or print retouching.

 

To achieve a good outcome, I have to work in small areas and the bigger the original file the better. The various functions of the Clone Tool can be used to both remove and rebuild areas of an image. The result is dependent upon the time taken (and of course the skill), but time and patience are very important.

 

Removing the Flash shadows was truly a 15 minute "rough" - kind of easy because the file was relatively small, but by the same token those results I posted are only good to a small enlargement size (maybe would be OK for a 5x7 inch print) and would probably not stand close scrutiny at a greater enlargement (e.g. greater than 10 inches on the long side) - but with a good copy file from a quality scan or a well executed copy image (using a camera and copy lens and copy lighting), I have had very good results from old prints, exceeding the "quality" (rough word for definition, clarity and acutance) of the original print at the same enlargement size: but that degree of removing and (especially) 'rebuilding' does take some dedicated time.

 

The technique of 'rebuilding' what was never there is also useful to fix Photographer error.

 

In the example below - in my haste to catch the train, I made the mistake of not framing all the house (top right corner is missing) and in addition to cleaning out all the "stuff" that I didn't like, I needed to rebuild that top right bit of the house and of course then rebuild all the other parts of the image underneath it.

 

The final product holds to good scrutiny at a 24" (long side) print.

 

Here is an A/B:

 

 

"Yellow House on the Corner" Austria 2014

(Fuji x100s)

 

WW

Edited July 25, 2017 by William Michael

[AlanKlein]

You can tweak it forever. But it won't make much difference to the child of parents since the love and remembrance comes through. Our brains adjust especially if the picture is of relatives. Pick the best, print and frame it, and put it up on your wall to enjoy.

[ross_lipman]

My submission. White level using the boarder. Decrease magenta. Increase overall saturation. USM. Crop. Still some work to do in cleaning up the background. Overall look retains a vintage feel, but brings colors back and increases detail retrieval.

 

[Kent Shafer]

Mainly, I used the "Clone" tool in Photoshop.

 

I have been told that there are other methods to achieve a similar outcome, but I learned how to use many aspects of the Clone Tool from the outset of my conversion to Digital Photography and thus, the Digital Darkroom processes. Photoshop was my first digital editor and I think that I often default to it because of that, often simulating a mimic of how I would have used film enlarging/printing or print retouching.

 

To achieve a good outcome, I have to work in small areas and the bigger the original file the better. The various functions of the Clone Tool can be used to both remove and rebuild areas of an image. The result is dependent upon the time taken (and of course the skill), but time and patience are very important.

 

Removing the Flash shadows was truly a 15 minute "rough" - kind of easy because the file was relatively small, but by the same token those results I posted are only good to a small enlargement size (maybe would be OK for a 5x7 inch print) and would probably not stand close scrutiny at a greater enlargement (e.g. greater than 10 inches on the long side) - but with a good copy file from a quality scan or a well executed copy image (using a camera and copy lens and copy lighting), I have had very good results from old prints, exceeding the "quality" (rough word for definition, clarity and acutance) of the original print at the same enlargement size: but that degree of removing and (especially) 'rebuilding' does take some dedicated time.

 

The technique of 'rebuilding' what was never there is also useful to fix Photographer error.

 

In the example below - in my haste to catch the train, I made the mistake of not framing all the house (top right corner is missing) and in addition to cleaning out all the "stuff" that I didn't like, I needed to rebuild that top right bit of the house and of course then rebuild all the other parts of the image underneath it.

 

The final product holds to good scrutiny at a 24" (long side) print.

 

Here is an A/B:

 

 

"Yellow House on the Corner" Austria 2014

(Fuji x100s)

 

WW

Sorry I didn't check back sooner—thanks for the help!

[William Michael]

You're welcome.