Painting from Photographs

Painting is difficult. Painting from photographs is even more difficult because cameras do not perceive the external world in the same way we do. There are similarities, to be sure, else we would not recognize photographs as replicas of that world. But there are significant differences between what a camera records and what we see. Experienced painters understand these differences and work around them in various ways. Student painters, particularly beginners, are often unaware of the differences and as a result produce unconvincing paintings when they attempt to copy exactly what they see in their photographs. This is true for both landscapes and people. (No one I know paints still lifes from photos.)

What are the main differences? The perception of values, color, and spatial relationships.

Values
Painters use the terms value or tone to describe the degree of lightness, or luminance, possessed by a given color or a specified object in a scene or image. Value range refers to the gamut of values found in a given scene or image. For convenience, painters have arbitrarily divided the full value range--which encompasses an infinite number of tones-- into a finite number of steps, usual five or ten. Some paintings, notably those of Caravaggio, Rembrandt, and other devotees of chiarascuro, have a broad value range, from the black of deep shadows to the near white of bright light.

Caravaggio, The Entombment (detail). As with most of
Caravaggio's paintings, this one has a broad value range,
from black to nearly pure white.

But most painters prefer a more limited range of values. Paintings in which lighter values predominate, as in Pissarro's below, are said to be high key.  

Pissarro, Hoarfrost. A high-key painting in which the darkest
darks are mere accents and the shadows are mid-range in value.

 Paintings in which darker tones predominate, as in Sargent's interior painting below, are--surprise!--said to be low key. 

 

Sargent, A Venetian Interior. A low-key painting in which most values
are in the low to middle range, and the lightest notes are confined
to small areas so arranged as to move the eye through the painting.

The term tonal range describes the range of values that a camera or other optical device is physically able to perceive under various lighting conditions. Today's digital cameras have far greater tonal ranges than film cameras but they still cannot capture the full range of values that the human visual system routinely perceives. Nor can they adjust as quickly or as subtly as our brains do to rapid changes in luminance. In all but the most intense light, we humans can make out details in deep shadows that can be nearly or mostly lost in photographs. We can also see details in brightly lit areas that are washed out in most photographs. 

Some correction in the darkroom or in software is possible, but the results, however otherwise satisfactory as images, still do not faithfully record what you actually were able to see in the field when you took the picture. The best you can hope for is the closest possible approximation.

Many digital cameras now feature an auto-bracketing feature where they take one shot using the camera's automatic settings for the scene, then take second and third shots one stop above and one stop below the standard. Using Photoshop or a comparable image editing program, you then can combine the three images to produce a single image that shows good detail and saturation throughout the value range. If you do a lot of painting from photographs, I highly recommend getting a camera that supports auto-bracketing, along with image-editing software that enables you to merge the exposures

The three following photographs were taken using the automatic bracketing feature on a Canon G12 camera. 

This image represents the Canon G12's automatic exposure
for this scene. The light areas are okay, but detail has been
lost in the shadows. Scenes of this type, with extreme value
contrast, are ideal for either automatic or manual bracketing.

Underexposing the image by one stop increases detail and color
saturation in the brightest area but loses all detail in the shadows.

Overexposing the image by one stop retains detail in the
shadows but washes out the bright area, which happens
to be the center of interest.

The following photograph was created using an image editor to merge the three preceding exposures above into a single image. The merged image retains detail in the shadows without washing out the bright center of interest. This is the image that should be used for a painting.

Color

Cameras do not render color in the way our visual system does. The algorithms, or set of rules, used by digital cameras to convert perceived light into colored pixels differ not only from those of our brain but among each other. And the best cameras, with the most sophisticated algorithms, are still incapable of matching the spectacular color sensitivity of the human visual system. 

All colors in nature are modified not only by the amount of light shining on them, but also by reflected lights bouncing off adjacent surfaces, creating a multihued radiance--an ambient glow--that a photograph cannot replicate. Cameras are capable of capturing the more obvious reflected lights, but not the myriad subtly nuanced reflections that alter the color of everything in small but gorgeous ways. 

Even if you are a professional photographer using the finest equipment available, you must deal with the built-in color biases that all digital cameras (and film, for that matter) have. Experienced photographers have done this for years, first in the darkroom, now in digital image editors. Even so, cameras, scanners, monitors, printers, and image editors all have color biases of their own, which is why our photos can look different when viewed on the monitor or printed. The solution is to synchronize all your devices so that they use the same standard for representing colors. This can be a hassle, but its a chore you definitely should do if you do want accurate color for purposes of entering shows, displaying on a website, or creating prints.

Spatial Relationships

Photographs do not convey accurate information about spatial relationships. Cues provided by binocular vision are totally lost. Other spatial information that people subconsciously absorb and process simply by being in a place is also lost. For example, atmospheric cues, such as the tendency of objects to become paler and cooler in hue as they recede into the distance, are rendered crudely. Using image editing software, you can adjust color balance, hue, and saturation to something very close to what you remember seeing. Auto-bracketing is a big help.

Depth of field is also a problem. If you shoot using a high f-stop, say f16, all edges in the photograph will be hard. This is also true of the automatic exposure settings on most cameras, which tend to be optimized for snapshots of people. Although some painters want all the edges, in their paintings to be hard, most painters prefer to restrict hard edges to their center of interest and modify all other edges to various degrees of softness. This accords with how we experience the world: where we focus our attention, the edges are crisp and well defined, while everything in our peripheral vision is to some degree softer. If you wish to reproduce this effect in your photos, you may need to switch from your camera's automatic exposure mode to its aperture priority mode, which enables you to use various lower f-stops in order to set the amount of soft background focus you prefer.

Parallax, the tendency for vertical lines at the edges of a photograph to curve inward, are a function of the curvature of the camera lens. Again, using image editing software, you can correct for parallax. 

Other Problems
Even with the finest images, painting from photographs still presents considerable difficulties for inexperienced painters. 

First, it is hard to re-create in the studio the intensity of feeling you experience in the field. The act of painting becomes more academic, an exercise in problem solving more than a creative response to your surroundings. This is why re-creating the spontaneity and freshness of a plein-air painting is so difficult in the studio. Even experienced professional painters have to work at it. Memory, of course, is an invaluable aid. The more you paint on location, the larger your database of remembered images. This is how experienced painters can create naturalistic lighting even when photos are not perfect, or can even change the lighting entirely from what is in the photo or was the case on the ground. There is no substitute for experience.

Second, it is too easy to become seduced by the photograph, to rely on it rather than on your personal sense of composition, color, mood, and other factors. The photo becomes a crutch that can weaken your work. You need to work hard not to copy its values and colors, not to take its composition as a given, not to reproduce every little detail it contains. 


Finally, painting exclusively from photos can represent a lack of commitment to the task of painting that may manifest itself in your work as preciousness, sentimentality, or banality. Painting from life, whether it be landscapes, still lifes, or figures, infuses your work with serious intent and authenticity. 

Some Suggestions

• Paint from life whenever you can. 
• If you must paint from photographs, make sure they are good photographs.
• Buy a good digital camera. A snapshot camera is adequate for most purposes, but not for the serious artist.
• Use your camera's viewfinder to compose your subject. Take photos from different angles and at various settings. Play with cropping photos to find new or more effective compositions.
• Learn to use image editing software to improve your photos and to experiment with various painting possibilities. More about this in a subsequent post.
• For studio paintings, the best source material is a combination of on-the-spot sketches or studies and multiple photos for use as reference.