Some of Ken's comments about human vision may well be correct, but his "15 feet" assertion seems to be quite wrong.† Is there scientific support for the claim that our brains recall people's facial features as they appear to be from about 15 feet away?
I suspect that we remember people from the distances at which we most commonly see them and interact with them, and that is generally not 15 feet.† Sometimes there are good reasons for photographing people at such a distance, but "how brains recall facial features" is not one of them, IMO.
That's a good question. I surmise that the distance of 15 ft might be a rough average of the many different distances at which we see people during our various activities, but I'm just guessing and I admit I have little experience in portraiture.
Doing a Google search to find the answer, I came across the following abstract of a scientific paper, which unfortunately one has to buy to read so I didn't get to read the full text, just the abstract below, which you might find interesting.
Realistic portraits, whether paintings or photographs, are traditionally obtained using perspective projection. Pictures of the face taken from different distances along the same viewing direction (e.g. frontal) may be scaled to occupy the same size on the image plane. However, such portraits differ systematically: e.g. when the center of projection (the camera) is closer to the face the nose is proportionally larger in the picture. These differences are small (for typical camera distances of 50-500cm): do they have an effect on how the face is perceived?
Ten naive subjects of both sexes, viewed equally scaled frontal pictures of 15 neutral-expression adult male faces, each photographed from distances of 56, 124 and 400cm. The photographs were corrected for lens distortion to obtain ideal perspective projections. The subjects were asked to rate each portrait according to 13 attributes (evil-good, repulsive-attractive, hostile-friendly, pushy-respectful, sad-happy, dishonest-honest, introvert-extrovert, violent-peaceful, dumb-smart, distant-approachable, evasive-candid, week-strong, unpleasant-pleasant). While the subjects were unaware of the manipulation, their ratings are systematically correlated with the distance: faces imaged from the closer distance appear significantly more benevolent (good, peaceful, pleasant, approachable), those taken from a larger distance appear more impressive (smarter, stronger). Intermediate-distance portraits appeared more attractive. The remaining attributes are not significantly different across distance.
Our findings suggest that painters and photographers may manipulate the emotional content of a portrait by choosing an appropriate viewing distance: e.g. a formal and official portrait may benefit from a distant viewpoint, while an effect of intimacy and opennes may be obtained with a close viewpoint. Multiple inconsistent viewpoints found in classical full-length portraits may be explained by the need to combine close-up views of some body parts, within an overall undistorted figure.
D. Freedberg, S. Shimojo, R. Adolphs, P. Hanrahan
According to this study, if you want to make your subjects appear smart and strong, take Ken Rockwell's advice and photograph them from 15 ft. (400cm in the experiment is actually a bit less at 13 ft, but let's not quibble.)
If you want your subjects looking as attractive as possible, get closer, and if you want them to appear as friendly and good-natured as possible, get even closer. Makes sense?
The link to the website is [a href=\"http://www.journalofvision.org/7/9/992/]http://www.journalofvision.org/7/9/992/[/url]