Racial bias in color film

Discussion in 'Casual Photo Conversations' started by cyanatic, Sep 19, 2015.

  1. Long before I took up photography in a "serious" way (whatever that means), I played in a blues band. Our drummer and bass player were both African American. I remember when we did a group pose for a publicity shot, the bass player didn't like the results and commented that "most photographers don't know how to photograph black people". Later, I came to think that it was mostly a question of how one adjusted exposure and color balance for darker skin tones.
    Regardless, I found these articles interesting and hope I am not duplicating another post (I did not find one).
    "Racial Bias in photography" --
    "Shirley cards and optimization of color film for Caucasian skin tones..."
  2. Is not whiteness really 'pinkness', Steve. Is blackness not a wide range of chroma / luminosity and light spectrum and light filament effects.? . Me, well lately I am middle gray:)....

    I once took some gigs shooting small weddings, what one would call budget weddings here. I knew at the get go that the mother of the bride, a local Filipina woman ,was not at all thrilled about her daughter choosing to marry an African American soldier. ( Shade of black was a class thing in the Phillipines) Got the proofs back to her. Her reaction: " Did you have to make him SO BLACK?" True. Can't recall what I did or said to the lab..Was and likely still is an exposure challenge at the least.

    I was just searching memories of popular films. Take the movie 'Spartacus' of 1960. Bare chested Kirk Douglas has a gladiator duel with a darker hued Ethiopian gladiator played by athlete Woody Strode. I wonder how that worked behind the camera.
  3. Steve, here is another thread that might be of interest.
  4. an a related subjct. I shot sdome e-3 ektachome, and blue jackets became green.
    this is a failure of that film to properly record color.
    a color film contains " color couplers: that when acted up on by the color development step
    convert to releas a certain color.
    years ago therer were many color failures . Like under flourescent lighting made things grreenish.
    it is hard for a film manufacturer to make these color couplers respond exactly.
    er asre fortunate they respond in a generally correct way.
    the human eye often sees diferently. perhaps differebt races or individuals see colors somewhat differently.
    my experience has been that many aftro-americans are recorder on film as a tan or yellowish color.
    I went to school with a pakistani who when weraring bathing trunks, had some flat black patcvhes on his chest.
    he complained of probls as the whites of his eyes were slightly tan.
    a frien from Ghana likely 100 percent african
    so things vary.
    what we perceve a black people are not necessarily black. our eyes may not be able to
    register the exact shade.
    and different people may not be able to
    distinguish small differences in color.
    my wife coud not determine shades of red ( yarn and cloth)
    she would ask me I have pale flecked blue eyes she had brown eyes.
    she was german and scots
    I am dutch irish and scots. so
    possible there are many other color vision differences in perception.
    in this case it is the limitations of making film.
  5. Interesting post Steve.
    I believe this issue is beyond just racial bias. It is always easy to politicize something that is more of a technical problem, and continues to be so. For me, this is about lighting, the nature of highlight and shadow, and perception of color. One of the components that makes a photograph successful is lighting. Natural lighting doesn't always do it, and most photos I see are done with natural light. This is why landscape photographers (great ones), wait for the right light, or know how to manipulate light to capture a scene. This also has to be done with people, to render highlights and shadows correctly. This, in many ways, is the division between serious amateurs/professionals and amateurs/hobbyist.
    One of the most discussed problems in photography is how to control highlights and shadows. With lighter skinned people, the separation of the whites of one's eye, and the skin. With darker skin (because all dark tones are not the same), allowing for more shadow detail, or lighting ratios.
    The last problem perception, has been the biggest problem/issue with people liking or disliking the results. Most people don't print their own work, and controlling color is quite difficult. Seeing color in a viewfinder is different than what film or a sensor sees, and part of becoming better as a photographer is the ability to see the way the sensor or film will render the scene. When color film became the rage, the big issue was, once printed the colors didn't look the same, and the main culprit was the person printing the images doesn't see color the way the shooter does. This led to transparency films (chromes), where the colors could be matched better, and better equipment was created to scan the chrome and match the colors.
    I shoot all races of people, for commercial and personal projects, and it hasn't been an issue for me because I very seldom depend on available light to capture what I'm shooting. I know there are many good photographers who achieve the same results through other methods (post processing, etc.). I just don't see images as a racial issue, as much as a photographic skill issue......
  6. It's surprising in discussions of this sort no one mentions how much contrast in the actual scene and how the standard toe/shoulder curve transfer function plays a part in this bias or perceived bias.
    Photography by its nature is affected by so many variables that I have a problem with seeing how there can be an intended racial bias to establish exposure standards as opposed to just overcoming the nature of the technology and compensating for it. Normal is meaningless in this sense due to all these variables.
    You can't have a good looking picture without contrast unless you're shooting high key lit scenes then everything is bright and washed out.
    The camera's metering has to fit so many various dynamic range sizes within one exposure which includes contrast appearance which can profoundly influence how bright individual elements such as dark colors must fall within the contrast range. Dark colors tend to stay dark no matter how bright you make the scene by increasing exposure. On top of that reality of how the scene actually looked is being compressed severely to fit within the viewing device's dynamic range be it a print or backlit display.
    Look at the example below of the leather belt next to pinky white Caucasian skin I just shot today. No matter if I lift shadows by decreasing contrast and/or increase exposure by comparison the belt appears to remain the same luminance and that is due to perception of how humans see overall contrast. The brighter the lightest elements the darker the shadows appear. That is a dynamic range issue, not exposure bias.
    Do "Normal" standards of exposure take this into consideration? I don't believe so. I don't even know where normal really falls within my own camera because its internal default metering is biased to not BLOW OUT HIGHLIGHTS and so makes even my skin look too dark.
  7. There was a time when Kodak color films were definitely optimized for pleasing reproduction of white flesh. Improvements came slowly but steadily over decades. In the 1960s surveys specifically looked at usage trends of different racial groups. In the 1970s internal pictures tests started using models with a variety of races. By the 1980s, the "UN scene" that included models of three different races was standard. In the 1990s racial equality finally appeared in some products. Kodak Portra films had flesh reproduction that was pleasing for all races.
    The tools to achieve unbiased color film were simple in theory:
    1. The color recording (spectral sensitivity) must be as close to the sensitivity of the human eye as possible.
    2. The reproduction of color (a combination of spectral dye densities and interlayer interimage effects) must be balanced to accurately reproduce all hues across the lightness range.
    3. Any attempt to produce "pleasing" rather than "accurate" color will introduce compromises that will make some colors (and some flesh tones) more pleasing than others.
    I have seen many studies of the color accuracy of color photographic products. I haven't seen similar studies of digital systems.
  8. Ron, did you actually see documentation that admits that Kodak color film processes and manufacturing were intentionally engineered to favor white flesh? I mean really!
    Since science was used to make this stuff, was it also used to determine pleasing vs accurate skin color of all races? Did the Kodak engineers look at the poor results of other skin tone reproductions and just throw up their hands and say f**k it, that's just the way it's going to be?
    What about the contrast curve that was applied in post processing. How did they decide or determine contrast reproduction?
  9. It's not just a problem with skin colour. Whenever I am photographed the film never captures my uncanny resemblance to George Clooney. Seriously, Agfa film for example, was balanced to reproduce skin tones under northern European (i.e. dull grey) skies. Kodak was balanced for West Coast light.
  10. More ignorance on how the photographic process records photons with a fixed contrast ratio established by fixed dark and light objects...
  11. Tim, there was no documentation. There were pictures that showed more pleasing reproduction of white flesh than black flesh. It wasn't an intentional move. It was a sin of omission.
    For a long time, brighter colors were preferred, even if they were exaggerated. Bright, oversaturated reds were easy to achieve. If reds are oversaturated, flesh is oversaturated. It was easy to compensate for white flesh--just bias the color balance a bit. The neutrals were on the cold side, but the flesh looked great. Black flesh is a more saturated color and received more amplification. A simple color balance change wouldn't fix it.
  12. Racial bias?

    All so-called "white people" do not look alike, nor do they have the same skin tones. Different ethnic communities here in the United States consider non-African people of various skin tones, i.e. Mediterranean, middle eastern, to be included in their definition of "white people." This also varies from the same ethnic community in different geographical locations within the U.S. My wife, who looked to be the whiteist of "white people", with blond hair and blue eyes, was raised by a couple in the U.S. who were then called "colored people," referring to what are generally called "black people" today. The gentleman toned very close to black and had non-reflective, more like flat black, skin. The lady was more a light chocolate brown, but to the dark side of caramel and seemed more glossy. Taking a picture of them with any film, or the early digital cameras, was a lighting problem. It seemed like a quarter stop either way could make one of them look unrealistic. Also, either the paper or the dyes in the paper would sometimes require a non-realistic tone to keep from an unflattering portrayal of their skin color.
    Olan Mills was an equal opportunity, lower cost, mass market, and very popular photo studio in the film days. A lot of their business depended upon people who lived locally near their offices. In the days of ethnic-centered communities, somehow using the same film their photographers and developing/printing technicians were able to turn out very nice looking portraits which were pleasing to the customers in their branch office's particular ethnic community.
    I'm of mixed race, but easily "pass" for white. Most people say I have a ruddy Irish tone to my skin. But then, knowing I have an Irish/Anglican last name, would they make some presumptions in their psychological perceptual screen? My brother (deceased) had what many people would describe as typical American Indian skin, eye, and hair coloring. Thinking about it, he looked more like what I would consider a Western/Plains Indian look, rather than the Cherokee of our descent. As we were growing up, nobody considered him anything but white. Could this be because they saw the family as being white, and therefore didn't notice the prominent American Indian features, which included nose, forehead and cheekbones as well as coloration?
    My best friend from junior high school on, is a mixture of middle-Italian and Sicilian-Italian. About twenty years ago, knowing I was turning more to my photography hobby in retirement, he brought me the negative and a print of him and his grown siblings having dinner with their parents. He asked if I could do something to make a better print, because to him, it looked like a nice African-American family out to dinner. It's not that he was afraid of being mistaken as an African-American, it just made his family look, in his eyes, peculiar, not what they really looked like. The problem was obvious. The "professional" photographer had taken a picture using flash in a dim restaurant, nicely bouncing it off a very odd colored "signature" tablecloth, producing a peculiar tint to the light that reflected off their skin and onto the film. He was so put off by it, he insisted on buying the negative so I was able to scan and photoshop out the result of the reflected light to where they looked a little more normal. Was the professional photographer, whose income depended upon photographing a wide ethnic variety of wealthy patrons in a locale of high-priced yacht clubs and restaurants, trying to racially give them the finger? I think it was just a lack of skill, marketing prowess, and the fact the person who took the picture did not do the developing/printing in their mobile step van. How did the guy in the step van know what they looked like? I think the photographer, showing them the result while she waited to get paid, just didn't notice the difference in a darkened restaurant with an old-fashioned incandescent bulb flashlight on the picture. It's human.
    In the late '90s, between the various film manufacturing companies, there were films marketed specifically for some non-lily-white or non-pinkish-white races or groups of people. There was a little family-run camera store where I saw, on their shelves, a space that had three different baby films (?). One row of boxes had what appeared to be a Caucasian-looking baby on the box. Another row of boxes had what appeared to be an Asian-looking baby. The third row of boxes had what appeared to be an African-looking baby. I wondered if the film in the boxes was any different.
    American Kodak made some of the first commercial digital sensors. But currently, most of them are made and designed in Asian areas. Are they making the sensors most flattering to Asian-toned skin? And, if so, what exactly is Asian-toned skin? And what tone and color combination within the various Asian groups are they specifically targeting?
    Come on now, folks. I think certainly with the last of the film emulsions and the current digital sensors, with filtering or photoshop, we should just do the best we can with what we have.
    That's life.
  13. Tim, there was no documentation. There were pictures that showed more pleasing reproduction of white flesh than black flesh. It wasn't an intentional move. It was a sin of omission.
    For a long time, brighter colors were preferred, even if they were exaggerated. Bright, oversaturated reds were easy to achieve. If reds are oversaturated, flesh is oversaturated. It was easy to compensate for white flesh--just bias the color balance a bit. The neutrals were on the cold side, but the flesh looked great. Black flesh is a more saturated color and received more amplification. A simple color balance change wouldn't fix it.​
    Ron, that's puzzling that you'ld associate saturation with darkness because I was ruminating on the color mixing functions of film dyes trying to sort out the lack of black or brown dyes in forming darker colors and was left with assuming the densities were made through the absence of light being reflected back from dark objects in the scene seeing that magenta, cyan and yellow dyes are much lighter colors.
    Now I've seen CMYK 3M Color Key transparencies used in commercial printing proofs whose dyes when viewed stacked on top of each other using halftone dot mixing function layered over paper looked pretty close to the print. Looking at each separate dye transparency would look much lighter than when stacked so the densities/HSL were fixed for viewing stacked. But black was included where as film dyes only use C,M,Y.
    So what is providing the densities for black and/or dark colors in film dyes? Absence of light reflected off subject in the scene?
  14. These kind of issues I have always taken seriously but I must say that I have never noticed or tuned into photos over the
    past 5 decades that bare out any notable bias that I can detect. I have not detected anything that would influence my
    perception or make me bias. From snapshots to media like Life Magazine to National Geographic I just don't see this.
    Film is a media that is not exact and anyone can find examples where they perceive tones or variations that are not
    perfect and no one can produce one that is a perfect reproduction be it skin color, jacket color or sky tones. For every
    statement I could say same the same about any picture of me or any subject. I believe the complaint is frivolous and
    motivated by a desire to address ones own desire to create an issue for their own agendas that are disingenuous or
    addressing their own inability to deal. From there the complaint becomes an indictment of the media, photographer and
    society. And of course one has to include ethnic publications such as Ebony etc. with these criticisms.
  15. Tim, there was no documentation. There were pictures that showed more pleasing reproduction of white flesh than black flesh. It wasn't an intentional move. It was a sin of omission.
    For a long time, brighter colors were preferred, even if they were exaggerated. Bright, oversaturated reds were easy to achieve. If reds are over saturated, flesh is over saturated. It was easy to compensate for white flesh--just bias the color balance a bit. The neutrals were on the cold side, but the flesh looked great. Black flesh is a more saturated color and received more amplification. A simple color balance change wouldn't fix it.​
    Well Kodak must not have had a problem reproducing African American skin tones in the '40's & '50's (the most racially charged era) going by these Kodachromes the most finickiest films to process correctly to get it to look right...
    So I don't know what to make of your explanation.
  16. That speaks to my point. Flesh varys ' so much even for "white skin" which include all kinds of variations from Sweedish
    to Irish to Slavic to Mediterian. Compare Obama's tone to his wife, who has the proper "black" skin? Can anyone say
    what comprises the definitive "ethnic skin".

    I used the term frivolous in the sense that if you want to apply the term bias there are real issues of "bias " regarding
    voting rights In this country and self perpetuating cycles of poverty that this issue detracts from and this alone could
    generate endless arguments ad nauseam that are not relavent to photography, but someone could still argue about
    neglect which would be endless with no consensus.

    With that said I am going to bow out of this subject that would not be resolved here , distracting and something that I
    regret commenting on itefrst place.
  17. Donald, I agree with your points, but I have to stress that science and how the photo chemical process functioned back then (not politics) at least from what I observed with a Staufer 10 tone stepwedge in the darkroom as a prepress technician in the '80's points to an issue that if both the exposure and the chemical processing isn't optimum on a consistent basis with consistent scene lighting it's far easier and likelier for darker colors to plunge into darker obscurity than lighter tones. IOW words there's a sweet spot with a narrow range of tolerance for getting it right. That's not a bias. That's not politics. It's just chemistry.
    Just examine the variation in contrast and overall darkness of other Kodachromes at that Shorpy site. They're all over the map. Some lighter, some overly darker. There's not a consistent one among them. Note in the low light factory worker shots how dark the backgrounds become from the front flash fall off.
    That shows where the problem is and what I observed in the stepwedge behavior during chemical development. It's not a gradually smooth behavior. It starts out slow and then it eventually gets quicker before I had to transfer to the stop bath or else the black threshold at tone 5 started bringing out highlight shadows that I wanted to be remain all white. This was high contrast line work for graphics. Continuous tone B&W into halftone dot had its own rate of gradualness and had to be watched like a hawk even more to control ink gain on the press or else suffer plugged up shadow detail.
  18. Tom, I thought that was the case also but could not speak to it because I am completely ignorant on these elements. I always had a sense that it was technical issues. The concerns that i could address was that it would morph into a socio-political issue that would distract from or muddy the waters of the real problems. Sorry I did say I would bow out
  19. Using a digital camera allows us to adjust settings if the images do not look "correct". More work is needed with film.
  20. There's another issue I've noticed recently with shadows and dark colors affected by contrast problems at my local movie theater using Sony 4K digital projectors in that certain movies made in the UK using Arri or Red Epic cameras have their gamma screwed up to where the black point becomes a milky haze and upper mids such as faces become really dim. The online trailers of these movies don't have this problem.
    "Woman In Black II" & "Man From Uncle" suffered from this to the point I couldn't see eyes and facial expressions in interior scenes and bright cloudless daylight looked as if it was overcast. African American faces were even worse at distinguishing facial features as they appeared concealed in the haze.
    With some online searching I suspect it might be the software used to transfer the finished digital movie to DCP standards for theater distribution, but I couldn't find hard evidence to prove this. Mostly DIY'ers were warning of this and not big movie production houses.
  21. Tim,
    I'm catching up on some of your comments. I was interested in the Kodachrome examples. I think the flesh reproduction of Satchel Paige is awful. I recall from the one time I saw him in person that he had dark brown flesh. It looked nothing like the magenta tinged flesh reproduction in this picture.
    The Army mechanic picture looks very good. The photographer did a great job of lighting. The background is underexposed enough to produce a saturated blue sky. The mechanics face is lighted to fill in shadows.
    The soldiers with flags photo has flesh reproduction that is not too bad. It is a little more red that it should be, but it's close. Part of this comes from the color balance adjustments in this reproduction of the Kodachrome transparency. Look at the white stripes in the flag. They are noticeably cyan.
    The 5 boys playing football looks good. The neutrals are neutral and the flesh is decent.
  22. The 5 boys playing football looks good. The neutrals are neutral and the flesh is decent.​
    That was the one that I'ld hoped would point to the lack of racial bias due to the fact that it's lit in broad daylight with no fill flash. So that proves even back then that the medium was capable of proper contrast and shadow rendering counter to the similarly lit shot in the Shirley Card Vox YouTube video I linked above.
    And if that video is referring to more modern consumer one hour photo lab type film, then the Shirley Card was of no interest if not rarely used by that market.
    The Satchel Paige image was only to show proper density of African American skin tone. Not all Kodachrome was as perfect as most shown in those Shorpy samples. What's of more concern IMO is the few numbers of African American subjects in that Shorpy Kodachrome collection. It took a long time rummaging through the collection to find the ones I linked above. Now that's racial bias.
  23. ".......few numbers of African American subjects in that Shorpy Kodachrome collection...... Now that's racial bias."
    Oh, really?
    I’ve seen the whole FSA Kodachrome collection along with many of their B&Ws. Did Shorpy pick and choose for their sample? In the collection as done by the FSA, people of all ethnic backgrounds were represented. Who was left out? The Wall Street Millionaires were, but then, was it prejudice or that they weren’t representative of the FSA overview assignment? Two other groups that were not represented much were Japanese and Germans. Was it racial bias against them or maybe the fact that the USA was at odds with them in WWII. Who wanted to parachute into Tokyo and/or Berlin to take their photos, not me!

    Back in my card playing days the deck had all sorts of cards in it. Today, it seems as they are all “race cards”.
  24. If the U.S.hadn't been built on slavery and continue to foster all kinds of institutional racism, that deck of cards you mention wouldn't be so stacked.
  25. I’ve never owned a slave. Anyone in the USA who is still living and is a former slave was not a slave in the USA, but was a slave elsewhere and is now free having reached the shores of the USA. I don’t think Fred G. owns/owned any slaves, but what do I know? I’ll let him address that.
    As to the USA being built on slavery, The USA did not start (and or build) slave owning on the North American continent, the USA ended it.
    I thought the “New World” was opened up by Europe to provide wealth for Europe by the French, Spanish, Portuguese, and British. THEY brought in slaves, mostly from Africa, but other places as well. THEY also killed, displaced and booted out the people living there on the North American Continent. Formerly oppressed subjects of those slave-owning countries, some slaves themselves and/or de-facto slaves, fought or bought their way into being the USA. The USA then did away with their inherited slave owning system. I don’t think that was a bad thing for the USA to do.
    Back to a photography point (if that is allowed) I still saw a good representation of ALL American culture, with the exception of the very wealthy urban segment, in the FSA’s collection of both Kodachrome and B&W pictures. Was that prejudicial against the Wall Street wealthy or just that The Farm Security Administration did not find much farm life in downtown Manhattan? I have not made a statistical analysis of Shorpy’s selection among that collection. I do not intend to do so either.
  26. Hey Art, I was referring to the anemic inclusion of African American subjects in the Shorpy Kodachrome gallery. It was meant as a general statement with a bit of hyperbole for emphasis.
    So you can put the flamethrower down, now. You've burned your point into my memory.
  27. I'll be happy to debate the issues of slavery and racism in some other forum, but I'll stick to casual photo conversations here.
    The racial bias in film was not active racism. It was more in the benign neglect category. The Kodachrome reproductions in Shorpy are not great examples of the racial bias in film because those early versions of Kodachrome do not have particularly saturated colors. I know this is counter to the Kodachrome myth, but by all objective measures, these early versions reproduced colors less saturated than the subject colors. When the saturation of consumer films was increased in the 70s and 80s the bias became more apparent and work commenced to offset the bias.

    If anyone has a K-25 slide that includes both white flesh and black flesh, if the white flesh is well balanced, the black flesh will be redder than it should be. Versions of Kodacolor, especially lower speed versions, will show a similar bias.

    Other manufacturers film had similar problems. I'm describing Kodak products because I'm far more familiar with their reproductions.
  28. Ron, I doubt that those selling the"racist film" story were concerned about reddish skin tones. The arguments that I see given are: 1) that
    dark skin is often printed too dark, and 2) that the so-called "Shirley card" uses mostly white female models. I think that the reddish skin
    tones are a non-issue with them. (Their other issues were with TV, etc.)

    I don't contest either of these, but don't think they're legitimate arguments.

    I know from experience that the first situation can be mostly handled by printing lighter, meaning that it's largely a lab issue. In the second
    case, I'm not sure what a "Shirley card" is, and I wish someone could explain it to me, as well as how they know this (please don't say you
    read it on the internet; at least say you read the instructions that came with the negative set). My experience includes time in a large lab
    which I would guess printed, inspected, and dust-spotted more paper volume than the entire active member base, at least in the forums, of
    photonet. Of course I have no idea what photonet members have produced, but I would be really surprised if they did more. I say this so
    people won't presume I'm naive about lab work.

    The sole purposes of printer control negatives, in my experience, are to 1) set up printer "slope" and 2) act as a bridge between video
    negative analyzers and printers. They in no way determine the color aims of a lab. In truth, the slope negs could probably just be some
    midtone grey, except that it would be hard to visually judge the significance of a color or density offset. A full range image is better for this.

    The business I was in did high volume portraits across the U.S. Not school pictures, meaning that we didn't have captive customers;
    everyone came to us by choice. We photographed pretty much every skin tone or complexion that existed in the U.S. For film, we
    considered just about every pro color neg film available, and never found any pressing reason to stray from Kodak's pro portrait films (in
    the lowest speed versions. Kodak was as good as it came, paired with the appropriate professional paper.

    Amateur films and papers are a different story, and with my much more limited experience, I don't want to say too much. But again, the
    quality of the processing lab is key.

    To Ron; the only thing we ever had problems with was the initial rollout of VPSIII. Initially the slightly higher contrast gave us problems with
    extremes of skin tone in the same scene, but early updates to the film handled this. Aside from this, Kodak doesn't have a thing to apologize for, at least in the pro color neg line,IMO.

Share This Page