Photoshop manipulations in journalism

January 14, 2004  |  Edward Tufte
32 Comment(s)

here

Topics: E.T.
Comments
  • Jeffrey Berg says:

    Photoshop manipulation used in this way is classic violation of graphical integrity. There is an almost measurable Lie Factor as you describe on page 57 of The Visual Display of Quantitative Information, Second Edition. Without a few whistle blowers, photo lies have become much harder to detect than your average info graphic fibs. A picture it seems is worth a thousand lies.

  • David Harrell says:

    Any thoughts on the “flipping” of a photograph so that a mirror image of the actual person/object appears?

    Relative to Photoshop alterations, it’s a trivial change, but as a guitar player I’m always annoyed to see pictures of well-known lefthanded musicians such as Paul McCartney and Jimi Hendrix playing their instruments “righthanded.”

    I’ve also noticed pictures in architecture and interior design magazines with both correct and mirror-image shots of the same room within an article. In most cases, it’s probably just sloppy work. But when done deliberately to “improve” the image I think it qualifies as a lie.

  • Mark Hineline says:

    This past summer, the sports editor at my wife’s newspaper ran a photo taken at a
    girls’ soccer game and removed the logo of a competing newspaper from a player’s
    jersey. After this was discovered, the newspaper bent over backwards to apologize for
    altering the photograph — with a story, a correction box on the first page of the
    sports section, and an editorial. The story link is below; it has both the correct and the
    altered images.

    http://nctimes.com/articles/2003/07/08/export388.txt

    The public response to this was very strange. Several readers wrote letters to the
    editor lambasting the paper for making such a fuss over so little, and taking the
    editor to task for publicly humiliating the photographer and sports editor who, the
    writer claimed, had shown “loyalty.”

  • David Glover says:

    Well…I actually think both the BYU and newspaper’s retouching is
    reasonable. Given that even I (an Australian) know that BYU is Mormon,
    and thus has fairly rigid “morality” codes, I’m no more surprised they
    edit tattoos and navels than that they’d crop a photo to cover a pair
    of breasts. While I don’t share their code, it’s not a secret, and
    this retouching is a logical consequence of it. I actually find it
    amusing and quaint.

    As for the newspaper, concealing its competition’s logo doesn’t alter
    the truth of the photo from a news perspective. It simply declines to
    provide advertising (which newspapers sell) for free. Would it have
    been inappropriate to, for instance, retouch out an obscenity (for
    instance the dutch word for “funeral”, which I understand is our
    beloved “f” word)? Or a t-shirt worn by a bystander that carries a
    racist message?

    I think we (as individuals and organisations) should be free to
    decline to promote other enterprises via their sponsorships. For
    example, Australia’s largest telecommunications company bought the
    naming rights to the Sydney Olympic Stadium. And on any commercial TV
    or radio station, newspaper etc it’s called “Telstra Stadium”. However
    our public broadcaster – which is mandated not to advertise —
    continues to call it “Sydney Olympic Stadium”. I, personally, do as
    well (until I get my slice of the sponsorship pie!).

    When a sponsor pays for the right to put their name on something, I
    don’t see any good reason for others, not bound by the sponsorship
    contract nor benefitting from it, to present their logo. And when the
    logo is a competitor’s…

    In short, I think these are two examples of reasonable image
    manipulation. I do think it’s appropriate for these (and all)
    publications to disclose their retouching policy and perhaps note that
    a particular image has been digitally altered in line with this
    policy.

    As for guitarists playing backwards, this is most likely simple
    ignorance (of both film emulsion and the musicians). Flopping rooms
    can sometimes be desirable to suit a page’s layout. In a formal
    architectural journal, it shouldn’t be done. In a Sunday supplement
    about decorating, I don’t think it’s a problem.

    Most media is skewed. All media is selective. And you don’t need to
    use Photoshop to present a misleading image. There are plenty of
    unflattering photos of celebrities and politicians that need no
    digital assistance to communicate a negative impression.

    I think it’s important for reputable news media to develop a policy
    and disclose it. I think to make that policy “never retouch” is
    absurd. An editor is mediating simply by choosing to run one story and
    not another; a photo editor is controlling the message by choosing a
    particular photo out of the dozens or hundreds provided. Technical
    flaws like colour biases, lens flare, messy backgrounds, uneven
    exposure etc that should be corrected. Removing a blemish will
    probably be appreciated in a

  • Mark Hineline says:

    While I find all of David Glover’s comments reasonable, I respectfully disagree.

    The important distinction that must be made is between retouching and falsifying.
    Can the distinction be codified? I don’t believe so. It is a matter of judgment.

    As far as I am concerned, altering photographs to remove meaningful representations
    (a tatoo or a competitor’s logo) is falsification. It is not mere editing. Cropping is
    editing; deciding to use a photograph or not to use it is editing. Using Photoshop to
    remove tatoos or logos is lying about reality. (Yes, I believe that there is a really real
    world, and that photographs are a dim mirror of that really real world.)

    What is the difference between removing tatoos in a photograph and writing the
    sentence “X, who unlike many players today has no tatoos on his arms, scored 12
    points on 18 attempts” when, as a matter of fact, X does have tatoos on his arms?

  • Gene Prescott says:

    It appears we are destined to encounter more “change and/or enhancement” rather than less if for not other reason than an increasing number of people have access to tools to accomplish whatever their personal notion of propriety is. So how to detect “authenticated” originals may be a more doable thing than keeping images from being altered.

  • Mike Combs says:

    In short, I think these are two examples of reasonable image manipulation. I do think it’s appropriate for these (and all) publications to disclose their retouching policy and perhaps note that a particular image has been digitally altered in line with this policy.

    Vendor identities, obscenities and other material can be masked, if necessary, with a low-resolution blur filter, or a good old-fashioned black box drawn over the offending material. This makes it clear to the viewer that the photo has been altered.

    A policy of disclosure is inadequate. Anybody remember the photo of Bush serving the turkey to the troops for Thanksgiving? Was the text disclosure, days later, that the turkey was a decoration, and not actually served to the troops, enough to set the record straight? I doubt a majority of people who saw the image also saw the disclosure.My point is that a picture sends a powerful message, and sometimes, its almost the only message that gets through.

    As a practical matter, tools like Google and Yahoo news photos separate photos from any surrounding editorial material. Disclosure of retouching must be made part of the image file itself.

  • John Morse says:

    I’m no scholar of Nederlandse, but should probably go ahead and dispel the earlier
    poster’s misinformation, even at the risk of veering off-topic.

    From the “Travlang.com” Dutch-English/English Dutch Dictionary:
    Funeral: 1. begrafenis, teraardebestelling. 2. begrafenis, graflegging,
    teraardebestelling.
    Funeral procession: 1. begrafenisstoet, rouwstoet.
    Funeral service: 1. rouwdienst.

    Of some interest (from the same source):
    Fokken: 1. breed, keep, raise, rear.
    Paardenfokkerij: 1. studfarm.

    So my guess is the etymology of our “f-bomb” has its roots in the ground in older
    germanic tongues and it probably took the Norman invasion to make it seem truly
    vulgar.

  • David Glover says:

    Thanks John, for setting me straight (I should know better than to use an unchecked
    anecdote!)

    Mike, your suggestions make sense, and would be a good part of a reputable media
    outlet’s policy.

    This is probably the challenge facing media that wants to have some credibility:
    making sure their policies are clear and any editorial decisions are transparent. The
    Web makes it so simple to check and compare with other sources. Personally I’ve
    lowered my opinion of a number of publications after quick Google searches revealed
    out-of-context quotes and loaded language.

  • Edward Tufte says:

    David Glover’s new message is a thoughtful and generous
    rethinking of his earlier arguments. Now and then on the
    internet, and I hope always at this site, there can be constructive
    exchanges of the sort seen in this thread.

    Much more on Photoshop constructions is found at our thread
    “Remarkable photographs from Mars [Oops! Fudged photos!]” :

    https://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0000X3&topic_id=1

  • Ziska Childs says:

    Fascinating…

    I cannot think of a single visual presentation which does not manipulate the
    facts and reality in order to “present” the argument of the presenter…. The very
    fact that you’ve translated something into 2 dimensions is- strictly speaking- “a
    lie”- you’ve
    removed time, smell, sound, and touch for starters….

    Let’s not kid ourselves about the integrity of the broadcast media machine…
    there is certainly integrity there… you do not pretend something is a live
    broadcast when it’s not … you verify your sources…etc…but if you have a
    choice of setting up the camera with bombs bursting behind you or a sylvan
    scene with ducks paddling in a pond… and you’re reporting a war- guess
    which camera angle you’re going to take…..Let me assure you the sensitivity
    of the audience to visual cues is not overlooked by broadcast journalism.

    Whether the Thanksgiving bird was eaten or not isn’t the point- more
    distressing to some of us was a blatant photo op- reported as “hard news”.
    This is a prevalent technique- especially in an election year. The line between
    reportage and photo op is certainly getting blurry but that has to do with the
    intent of the presenter-not the integrity of the image (why?- because there *is*
    no image on your evening news which is a “true” image- sorry guys- it just
    doesn’t exist).

    One of my favorite publications is (which is in danger of extinction) “The
    Journal of Decorative and Propaganda Arts” published by the Wolfson
    Foundation. Not only great “eye candy” but a very good read.

    BTW- there is an interesting story about the origins of the “f” word which
    involves the Royal Navy… but that would not be appropriate in this forum. It is
    an acronym.

    Ziska Childs

  • Mark Hineline says:

    I seem to be enforcing a rear-guard action against overzealous relativism
    here, but I feel the need to critique Ziska’s opening argument. Yes in some
    ways two-dimensional media are “lies” about reality. The world is indeed not
    flat, and much that would impinge on our perceptual faculties is left out of a
    photograph.

    But if a photograph is a “lie” in this sense, it is a conventionalized lie. We learn
    how to interpret or “read” photographs (and video). We can also improve our
    ability to read photographs — to become better and better-trained perceivers
    of such images.

    As a part-time woodworker, for instance, I am always interested to look a
    photographs of interiors in such periodicals as “Dwell.” I am always aware
    that photographs of interiors are long exposures (in most cases). I find myself
    wondering what the “real” color of a piece of furniture is.

    None of this is to the point about either the BYU or North County Times
    examples. In these cases, messages were intentionally removed from
    images. No matter how trained a reader one is, it is impossible to reconstruct
    the elided messages in the retouched photos.

    Therein is a distinction with a difference.

  • Dave Nash says:

    Yes let’s not carry this relativism too far and have it decay into the argument from illusion. It is a large jump from selecting, or even cropping, actual photographs of reality to manipulating photographs into non-reality. The facts and motivation of BYU’s photo manipulation are not in doubt: from the Salt Lake Tribune we learn that the BYU Associate Athletic Director for Communications “said it’s easier to doctor photos than deal with irate callers demanding to know why athletes are being immodest or wearing tattoos” (http://tv.ksl.com/index.php?nid=5&sid=69483). I suggest that in this case the “honor code” is really the “donor code”.

    While technology now allows anyone to alter photographs it also allows anyone else to raise a worldwide alarm when it happens. The misrepresentation of a college athlete’s flesh is nearly irrelevant, but even the US Government has been checked:

    “In September 1990, the first Bush administration said that 265,000 Iraqi troops and at least 1,500 tanks were in Kuwait, many positioned along the border with Saudi Arabia, poised for an invasion. Based on those reports, the Saudis asked a coalition of forces, including the United States, to provide protection.

    The St. Petersburg Times purchased Soviet commercial satellite photos of the border and had them examined in detail by two imaging specialists, including Zimmerman. They found no evidence of significant military buildups along the Saudi border.

    After the war, Colin Powell, then chairman of the Joint Chiefs of Staff, now secretary of state, acknowledged that there could have been far fewer Iraqi troops in Kuwait than originally believed.”

    (http://www.globalsecurity.org/org/news/2003/030315-satellites01.htm)

  • George Klucsarits says:

    This past Saturday I attended the Nikon School, a one-day seminar on
    photography, in Atlanta. One of the instructors, Fred Sisson, a veteran
    newspaper photographer, related how one of his editors would “manipulate”
    images of political rallies depending on how he felt about the particular party.

    If the room was sparsely populated, and it was not his favored party, then the
    photo would emphasize the many empty chairs. On the other hand, a similarly
    attended rally of a favored party would be photographed to create the
    impression that many more people were present.

    By taking advantage of cropping, point-of-view, selective focusing, etc,
    different photographers can easily present different versions of “reality”
    without resorting to any additional manipulation, either in the darkroom or the
    computer.

  • Mark Hineline says:

    The best example of editing in camera is the canonical photograph of the driving of
    the golden spike in Promentory, Utah, at the completion of the first transcontinental
    railway (in the United States). For this famous photograph, the coolies who had just
    laid the last bit of track were moved out of the way and replaced entirely with Anglos.

    There was another famous example of this, and it raised the question of which is
    more “real:” the arrested image, or moving picture. This was a shot of George Schultz
    testifying before a Senate committee (during the Nixon administration, if I remember
    correctly). The arrested image, which ran on A1 in papers across the country,
    captured what looked like a moment of deep shame, with Schultz’s hand on his face.
    Video of the same shot shows that Schultz was in fact rubbing his face. The emotion
    portrayed in the video was clearly boredom, not shame.

    This distinction between the realities of arrested image versus moving image is a
    significant aspect of the criminal trial in the Rodney King case. The policemen’s
    defense attorneys reduced the well-known video of the beating to a series of arrested
    images, thus changing not only the interpretation of reality but the reality itself.

    So yes, photographic images can and do manipulate reality. That does not make such
    manipulations any less deplorable. I see no reason to throw up our collective hands
    and say “all is fiction.”

  • Mark Hineline says:

    Having just written the above, I sense the need to point out that I do not argue for a
    naive realism in visual representation. I am aware of the countless ways that arrested
    images have been manipulated; in fact, I wrote a chapter of my dissertation on this
    matter ten years ago (relative to specimen images of fossils, which I call “proxies.”)

    Thus, I am not arguing above that photographs should not be manipulated using
    digital technologies like Photoshop because photographs pre-Photoshop captured the
    truth and nothing but the truth. I am, rather, arguing that such manipulations should
    always endeavor to increase visual truth rather than to diminish it. There is no
    question in my mind that digitally removing tatoos from a basketball player’s arm
    because of the consequences that the discovery of the tatoos would have is an
    example of a manipulation that diminishes truth. It seems to me that this is not
    arguable, no matter how many examples one profers of prior manipulations of
    images, digital or pre-digital.

    Falsifying does not justify falsification.

  • Ziska Childs says:

    First I must apologize for the use of incendiary absolutes. I am
    normally speaking with people who have little or no visual vocabulary.

    “much that would impinge on our perceptual faculties is left out of
    the photograph”.. Quite. Remember the marvelous scene in Amadeus? “Too
    many notes!” (A typical and telling executive response)

    Thank you George for an on target illustration.

    The problem with photographs is *precisely* because they are “a
    conventionalized lie”. They are frequently used as the MacDonalds
    french fry of visualinformation. Aside: Embedded photographs within a
    presentation can be a good adjunct but they will not tell you what
    kind of joins are used in the cabinet (I’ll wander farther off track
    to another terrific book for cabinet makers “Traite D’Ebenisterie” by
    Lucien Chanson editions H. Vial- afraid my copy is too early for an
    ISBN number)

    What color is it really? I ain’t touchin’ that one….

    Back to the tattoos…The choices one makes determine the viewer’s
    reaction. It doesn’t matter if you are creating a photograph, a
    drafting, or a graphic Those choices may be motivated by many things.
    They are almost always driven by a certain task (show the number of
    individual deaths by cholera by location), a desired result (often
    crossing the line into propaganda) , loss of “face” (“Erase that
    tattoo for ol’ BYU!”), or they may be subconscious. The viewers
    reaction will almost certainly be motivated on a subconscious level.
    Whether that manipulation is by digital means or setting the point of
    view or cropping the image is relevant IMO only as to the motivation
    behind the manipulation (how successful that manipulation-if it ties
    into the reptilian part of our brains- which “point of view” normally
    does- it tends to be *very* successful)

    To quote another thread …one must start with “an idea”. That “good
    idea” will determine the form of the solution.It helps if the person
    presenting the idea has a touch of integr

  • Ziska Childs says:

    What if the image of a basketball player had been a pen and ink drawing (a la
    Wall Street Journal)?

  • Mark Hineline says:

    I was wondering the same thing. Asking this question actually raises a number of
    issues that have been obscured in the discussion thus far. Most important issue: what
    is the relationship between subject matter and technique in executing a particular
    visual representation.

    In the case of sports photography, the primary goal is to stop the action and to
    provide a crystal clear image of an athlete in the midst of execution. And there’s a
    good example of a “conventional lie:” note that this one goes back as far as
    Muybridge. To get that kind of image, a lot else is sacrificed, especially depth of field.
    The idea is to capture muscle tone and tension, beads of sweat, and the concentration
    on the player’s face.

    What would count as comparable conventions for pen and ink? Quite different, I
    should think. Not at all like anything the Wall Street Journal does. In the case of pen
    and ink, detail is sacrificed for what art instructors call “gesture:” a sense of
    movement. And in this case, the tatoos would not appear.

  • Melissa Spore says:

    The seriousness of manipulating digital images was made clear by Angela
    Murphy of The National Research Council (Canada). An article in this week’s
    On-Campus News at the University of Saskatchewan reported her talk in the
    Canadian Journal of Microbiology Seminar Series.

    Beyond outright falsification,Murphy said, there are currently two schools of
    thought about manipulating digital images in the sciences. The first argues
    “that any global adjustment to the electronic image is okay, and that any
    specific alteration that fouces on a particulat area is not…The second school
    is less stringent, allowing for specific alterations as long as they don’t obscure
    the main points of the research.”

    Murphy demonstrated problems with the first position through a digital image
    of a stained lung carcinoma. The colour intensity had been altered to make
    the image more vivid, but the result suggests a stronger immunological
    reaction than occurred.

    For the second position, she examined a situation in which healthy cells were
    copied and reintroduced to cover up errant knife marks in an image of a feline
    brain. “In the process, however, the cerebral cortex was accidentally filled in,
    rendering the data unusable.” These are only two examples, which Ms.
    Murphy has collected in her role as managing editor of biomedical journals for
    the NRC.

    Her suggestion is to document any alterations, including technique and
    software. This documentation is then attached to and accompanies the
    image.

    She spoke approvingly about using scientific images as works of art, for
    journal covers, etc. “But careful attention needs to be paid to graphics that
    represent data.” This echoes ET’s comments in the Larkin’s Twig thread (and
    elsewhere). We have an obligation to the data.

    Melissa

  • Ziska Childs says:

    Melissa,

    LOL!- well who needs that pesky cerebral cortex anyway?

    Mark,

    Pen and ink conventions allow you to do much more than remove the tattoos –
    you can also (easily) obscure the ethnic origins of the athlete.

    Journalistic ethics require more than adherence to artistic conventions.
    Airbrushing a tattoo is IMO a pathetic attempt to break that journalistic moral
    code.

    Journalism should convey both immediacy and honesty without showing the
    bias of the reporter or the photojournalist. In order to appear trustworthy
    broadcast and print media use a large bag of visual tricks. The textbook
    example of this is 60 minutes with the two viewpoints used for the interview
    camera the medium shot for the interviewer and the tight shot for the
    interviewee.

  • Cris says:

    Although the comments regarding integrity in a journalistic context are spot-on, is that the right model to judge a media guide? Such a guide is essentially a marketing tool, and the fact that photos in it have been retouched should not really be any more surprising than if it was the press kit for a movie.

  • Ziska Childs says:

    The marketing of journalism is not an unrelated subject…. But in answer to your question I believe that if you’re falsifying the photos (using someone else’s body for instance) in order to make the fortify the schools “honor code” then there is something seriously wrong with the way in which that honor code is upheld.

    The implications of this may not seem as dangerous as erasing the cerebral cortex in medical illustrations however I will stand firm in my view of photojournalism being a particularly powerful and dangerous tool for manipulating public opinion. A tool which should be handled with extreme caution.

    Recently I participated in college entrance interviews. One of our stock questions was “Describe a moral ethical dilemma which effected you personally.” The most interesting answer was from a High School Senior who was the chief editor of her paper. The story was about students reaction a mandatory pledge of allegiance at the start of every school day. Her photographers brought back a picture which included two students giving a “Nazi salute” Her moral dilemma was whether or not to crop these two kids out of the photo. Why? Because her mother was the teacher in that classroom and she knew it would get her mom in a lot of hot water.

    http://www.thedenverchannel.com/news/2563305/detail.html

    http://www.issues-views.com/index.php/sect/23000/article/23072

    Plan would let students opt out of pledge

    (Here’s what surprises me… that this story didn’t get picked up by the National News.)

    [link updated March 2005]

  • Eric Wright says:

    A quote from Visual Explanations (p. 25) is relevant here [emphasis mine]:

    Recently, inexpensive computing and ingenious techniques for image processing have provided endless new opportunities for mischief. Arbitrary, transient, on-sided, fractured, undocumented materials have become the great predicament of image making and processing. How are we to assess the integrity of visual evidence? What ethical standard are to be observed in the production of such images? One way to enforce some standard of truth-telling to insist that the innocent, unprocessed, natural image be shown along with the manipulated image, and further, that the manipulators and their methods be identified.

    This seems like a fair strategy. If BYU wants to remove the tattoos, I think that comes at the price of mentioning something like “The above photo has been altered to remove tattoos on Mr. Araujo’s arm.” Anything less is dishonest.

  • Michael Round says:

    Requiring “before and after” documentation surely would decelerate the image-doctoring in the first place, as few people, it seems, would like it known such practices do take place. Likely, a better shot would be found, or simply the “before” image would rule out.

    A larger part of the reporting process, however, is why the particular image was shot in the first place. Of all images a reporter could submit, why focus on the French President in a boat (to use the recent photo-doctoring incident as an example)? Does submitting the undoctored image “move us closer to the truth”? Why did the reporter claim this to be news in the first place?

    War coverage is a further example. Iraq is a big country. Reporters cover “what’s happening” and submit images. Undoctored images don’t necessarily mean “truth”, as this is but a small part of “the truth, the whole truth, and nothing but the truth”.

    What is one to do, then?

    Walter Shewhart’s first rule for the presentation of data, taken from Donald Wheeler’s book, Understanding Variation, assists in this regard:

    “Data should always be prsented in such a way that preserves the evidence in the data for all the predictions that might be made from the data”

    Data … words … images … As Dr. Tufte says, “it’s all evidence”.

    Michael Round

  • Art LaMan, III says:

    I saw this short article in Technology Review today:

    http://www.technologyreview.com/Infotech/20423/

    Research into techniques to identify manipulated photos through analysis of lighting.

    Art LaMan, III

  • Marshall says:

    For a lighthearted look at this, there is the blog PhotoShop
    Disasters
    , which points out bad image manipulations in advertisements, book covers, etc.

    My current favorite is Castles in the
    Air
    (count the hands)

  • Christopher Wilson says:

    Great discussion, but what no one has mentioned yet is why does BYU choose to “take the easy way out” and
    manipulate these photos?

    It seems to me that if BYU has a problem with immodesty and tatoos, they should deal with the actual bodies (the
    students) themselves, not their representations. If the institution has such high moral standards, then why do they
    have students with bare midriffs and tatoos? They are probably asking themselves this same question, hence the
    Photoshop manipulation to “solve” the problem – whereas they should really be addressing this internal
    inconsistency . . .

  • Will Oswald says:

    Interesting article on the New York Times website. It appears that Iranian state media manipulated photos of the recent missile launch to cover the picture of what could be a failed missile launch. What is unfortunately not described is how the NY Times discovered this – was this a keen-eyed staffer, or some automated approach to catch such issues…?
    http://thelede.blogs.nytimes.com/2008/07/10/in-an-iranian-image-a-missile-too-many/index.html?hp

  • Edward Tufte says:

    Intriguing discussion and interview by Errol Morris here

  • Will Oswald says:

    Yet more state photo manipulation, this time seemingly from North Korea. As the BBC reports, a recently published photo of Kim Jong-Il at a military parade has several inconsistencies that suggest his image was inserted to dispute rumours of his ill health, or possibly even death.

  • Prem Thomas says:

    Evidence consumers take note:

    Adobe Photoshop CS5 gains new powers, opens new frontiers in image manipulation.

    See http://cs5.org/?p=147

Contribute

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.