Imagery is powerful. Ninety percent of all information that comes to our brain is visual. Thus, our brains spend more energy processing visual information than information from all other senses combined. It’s probably no surprise, then, that we process visual information much faster than text.
Fake imagery can be powerfully deceptive. We grow up expecting that what we see is true. Thus, if what we see is fake, we develop false understandings. Models photoshopped to be thinner and more attractive have set unrealistic standards of beauty. And sometimes, we purchase products only to discover what we purchased is not what we saw.
The Difference Between Image Processing and Image Manipulation
Some image manipulation is minor and misleads us in insignificant ways. In fact, we routinely make modifications to a photograph as we prepare it for publication or print. We adjust colour, contrast, and white balance in order to correct for equipment inadequacies and to recreate visually what we saw and experienced when we took the photograph.
We also routinely remove or minimize imperfections in a photograph. When we look around us at people in real life, what we see is influenced by the movement around us and in peoples’ faces. Our focus is also affected by our shifting visual attention. We don’t see details as much as we see environments, faces, and expressions. A still photograph doesn’t offer us those distractions so our eyes notice dust spots, imperfections, and blemishes that we might otherwise not have noticed. Subtle retouching reduces those distractions, approximating our real world appearance.
Image Manipulation for Art and Advertising
Some image manipulation is radical, done for artistic or commercial effect. The manipulation is obvious, creative, and beautifully done. These post-production artists are widely regarded and respected. Christophe Gilbert is one of my favourites. He’s especially known for his ability to manipulate paint and water, but I like Gilbert for his wit. He is, understandably, in high demand in the advertising sector.
Changing History With Image Manipulation
Some image manipulation alters our understanding of current events and changes our perception of history. For example, Joseph Stalin, the Soviet Union dictator, developed a reputation for altering photographs to remove people who had become his enemies. In the photograph below, Nikolai Yezhov is visible at Stalin’s side in the original photograph (right). After Yezhov fell out of favour with Stalin, Yezhov was removed from the photograph (left) and from Russian history.
In 1950, at the height of an election campaign, two photographs were composited to place an American senator (on the right) in conversation with a leader of the American Communist party (on the left). Given the anti-Communist mood in the United States in the 1950s, it’s probable the fake photograph contributed to the senator’s electoral defeat.
Because image manipulation can have such wide-reaching implications, the rules governing image manipulation in journalism are fierce. The following example of rules in photojournalism is from the New York Times:
Images in our pages, in the paper or on the Web, that purport to depict reality must be genuine in every way. No people or objects may be added, rearranged, reversed, distorted or removed from a scene (except for the recognized practice of cropping to omit extraneous outer portions).
Still, some photojournalists are tempted to fake their images. In 2003, Brian Walski, a well-known and respected journalist for the Los Angeles Times, was fired after submitting a composite of two photographs taken in Iraq. A newspaper employee noticed that some Iraqi civilians in the background appeared twice, providing the clue that the final photograph was a composite. Walski acknowledged his editing and admitted that he had made a slip in judgement that he couldn’t explain even to himself. He accepted responsibility for his actions and subsequent firing.
In 2010, World Press Photo disqualified Stepan Rudik’s photograph after learning that the photographer had removed the tip of a boot from the background. The boot added nothing to the photograph and removing it didn’t change the story. However, World Press Photo strictly applied the rule that “[t]he content of the image must not be altered.”
Deception With Image Manipulation
Sadly, some radical image manipulation is done with intent to deceive and even to profit from that deceit. The practice is widespread in social media, luring millions of followers to view and unwittingly share faked photographs. One Twitter feed has almost a million followers and is estimated to generate tens of thousands of dollars a month in advertising revenue. This account, like so many others, shares uncredited images and images that have been faked either by attaching inaccurate or misleading captions to historical photographs, or by deliberately and extensively manipulating the image.
The practice of sharing stolen and faked images is now so widespread that individuals are launching social media accounts dedicated to “outing” the offenders. Paulo Ordoveza is one of these champions, running the Twitter account @PicPedant. Ordoveza puts in hundreds of hours using Internet resources to research the provenance of photographs, identifying the original creator of uncredited photographs and exposing the manipulations in fake images. Antiviral is another of several similar social media feeds.
How to Spot Fake Images
Verifying and certifying photographs as authentic is big business in the photography industry, especially for photojournalists, media editors, competition judges, police services, and courts. Online and custom services analyze an image’s metadata, using complex and evolved algorithms to detect changes. Pixels can also be mapped and filters applied to images to reveal alterations. Another method known as “JPG ghost detection” analyzes a file to see if the image has been saved multiple times at different quality settings — a sign that the image has likely been manipulated.
This type of image forensics is valuable and has its place, but it’s not practical or affordable for assessing the hundreds of images you are likely to see over a week or month on social media, in portfolios, or just generally scattered on the web. You can, however, train your eyes to detect image manipulations. This kind of awareness training will also make you a better photographer.
To begin your awareness training, start with some of the following techniques.
Take a mental step back and look at the image as a whole. Could the scene in the image really have happened the way it is shown? Does something about the image look too spectacular to be believable? If the scene doesn’t seem plausible or the image is just too fabulous, the image has probably been faked.
Perspective & Proportions
While taking a broad look at the image, assess the perspective and proportions in the image. Does the perspective seem realistic, or does the perspective shift somewhere in the image? If it does, it’s likely the image is a composite of two or more images taken from different angles. Variations in perspective, lens length, and depth of field will also show a difference if not expertly adjusted.
Also consider whether the various components of the image seem to be in proportion with one another. For example, does a person’s head seem to be too big for their body, or their feet too small for their height? Does furniture seem overly large or windows too small? Any irregularity in proportion is a clue that suggests manipulation.
Lighting irregularities or discrepancies are easy clues to spot once you know what to look for. Begin by following the trail of light from its source in the image and determine whether the highlights and shadows are moving in the correct direction. For example, light coming from the left side of an image will create shadows falling to the right. All shadows—even those under noses and eyes—should fall in the same direction. Reflections, like shadows, should fall in the same direction.
Highlights should be in the light’s path or on a reflective surface struck by the light. Also evaluate the highlights for realism. Are they a little too bright or perhaps there’s one too many?
Shadows and reflections should also be realistic in location, size, and depth. For example, a strong light source from above will create a deep shadow beneath the subject. A studio light in a softbox placed in front of a subject on a 45 degree angle will produce a small, soft shadow behind the subject on the side opposite the light. A lack of shadows can also be a clue that the lighting has been manipulated in the image.
Textures and Patterns
Textures, patterns, and objects rarely repeat identically, especially in nature. Textures that appear to be nearly identical or repeat oddly — for example, a dark patch of grass that appears over and over again — are clues that an image has been altered. Similarly, if a dog has three identical spots or a clouds in a stack looks suspiciously similar, chances are the image has been manipulated. Duplicated trees, branch patterns, bushes, or flowers are also signs of image manipulation. A quick way to find these repetitions is to cross your eyes as you look at the image so duplicates will stand out more.
You can also detect image manipulation by looking for irregularities in textures and patterns caused by the physics of photography. For example, textures take on even more of an organic or irregular look in photographs because light creates shadows and depth, which are enhanced when compressed into a two-dimensional image. Patterns should have converging or widening lines, depending upon the perspective in the image and the lens length used. If these photographic changes are absent or inconsistent, the image has probably been manipulated in some way.
Irregular or broken lines in patterns also reveal evidence of image manipulation. It’s a tricky task to align patterns and keep perspective when cloning or compositing. Subtle changes along pattern edges and lines, irregular or broken edges, or slight directional changes in lines are all evidence of image manipulation. Image manipulation can also result in curves or waves in lines, or irregular changes in curves.
Scraps and Blurs
Selecting objects in an image is a demanding task. Volumes have been written describing different methods for making clean selections. Look closely in any but the best edited image and you are likely to find scraps, irregularities, or halos around edges. Depending upon how well an image has been modified, a piece of background may even tag along when an object has been copied and pasted into either the same or a different image.
Edges and sections that have been manipulated may also appear subtly over-blurred. Alternatively, edges that have been manipulated may appear overly distinct from the background, sometimes appearing as if the object is sitting on a layer - resting on, rather than integrated with, the background. A lack of shading around the edges of objects is another indication that an image has been manipulated.
Yet one more clue that an image has been heavily manipulated is the appearance of circular patterns in large areas of an image. The circular patterns are often caused by the round shape of digital tools used for image manipulation. For example, circular patterns in a sky, road, or wall are often a sign of cloning or healing.
It’s standard to retouch portraits, usually with the intent of enhancing the subject’s best features and minimizing their worst. In fashion and magazine photography, portraits — and, indeed, whole bodies — are routinely retouched, reshaped, smoothed, and polished. Because we see these manipulations so commonly, we’ve come to accept them as normal or at least not unusual.
Conversations about the ethics of manipulating bodies in images are extensive. It’s beyond this article to explore that topic in depth. However, with attention and practice, you can begin to distinguish between retouching and heavier manipulation.
Begin to look for evidence of extensive image manipulation in a portrait by examining skin colour. Skin colour that appears very consistent, polished, or too perfect in colour suggests the image has been edited. Skin colour can also appear a bit off if it’s been manipulated — perhaps too tanned, too fair, or too pink.
When assessing a portrait for manipulation, also look for skin with very little texture, too few wrinkles (even children will have a few “crinkle lines” beside smiling eyes), and little variation in colour and tone. Blurred edges — around the jawline, for example — are another sign of image manipulation. Also evaluate the shadows on a subject’s face: are they realistic and proportionate given the light source and the person’s face shape?
Eyes and teeth offer additional clues of manipulation. All eyes have little veins and almost all eyes and teeth have irregular surfaces. Eyes and teeth that are too white or are missing subtle variations in tone have likely been manipulated. Highlights in the eyes should come from the same direction, and the number of highlights should match the number of light sources in the photograph. Eye colour that is too bright or too perfectly uniform is another indication of manipulation.
Colours and Tones
After looking at an image in detail, take a step back to take another long look, this time to assess colour and tone. Tones usually show subtle gradations and change as they move from dark to light. Colour balance should be consistent if the light source is uniform. For example, an object with a cool colour balance in the middle of a warm-balanced photograph is likely an addition from another photograph. Changes in saturation or unrealistic saturations are two more clues of manipulation.
Assessing colours in the shadows can also tell you a great deal about an image. The colours in the shadows should align with with the colours in the highlights. Colours in the shadows will be darker, certainly, but should show a coherent white balance and tone. Colours in the shadows should also appear slightly desaturated.
If you are assessing the digital file of an image, try boosting the overall saturation of the image. Tonal and colour differences in the photo will be easier to detect. Variances in colour noise will also become more obvious.
It is easy, tempting, and sometimes justified to manipulate a photograph to obtain a perfect image. Often, careful processing and a few touch-ups can transform a mediocre image into something special. Other times, extensive retouching and changes can result in an artistic photograph that is not meant to reflect reality as much as it is meant to convey an emotion or message. Sometimes though, manipulating an image can amount to dishonesty based on a deliberate intent to deceive. Knowing what to look for to determine whether an image has been manipulated both prepares you as a viewer to properly evaluate an image and as a photographer, to improve your own editing techniques.