That moon is fake

The era of digital imaging seems to have resulted in increased skepticism regarding "manipulation" or "digital trickery" in photos containing the moon. Fact is, nearly all photos containing both a full moon and any earthly features - now and since the dawn of photography - have been manipulated extensively. One reason is that the moon is much brighter than any landscape when a full moon is visible. The moon is always lit by full sunlight (unless the earth is in the way), so a good estimation of exposure requirements comes from the "sunny-16" rule, i.e., f/16 at a shutter speed of one over the ISO rating. Earth-moon scenes within a day or two of a full moon are simply too dark for film or digital sensors to record both at the same time. The possibility of properly exposing the moon and a landscape in the same shot increases as you back off from a full phase of the moon. The earlier in the phase, the more sunlight is hitting your landscape. Very rarely, you might catch a moon near the horizon, two days before it's full, with enough airborne dirt or industrial pollution to darken it while not hopelessly blurring it. Three or four days before full, you might post-process a single exposure and end up with a winner.

Ansel Adams post-processed photo-chemically in a darkroom, greatly varying exposure of different parts of a print by dodging (blocking parts of the image with a piece of cardboard) and burning (darkening the print by aiming the enlarger head at a portion of the image). This technique resulted in Moonrise, Hernandez, New Mexico, 1941, his most famous photo, and one of the most famous photos of all time.

The phase of the moon and its position above the horizon reveal that this wonderfully moody night scene was actually recorded in late-afternoon sunlight. He reported using a Wratten No.15 (G) filter with presumably panchromatic ISO 64 film, exposed for one second at f/32. Adams didn't record the time or date - it was somewhere in the 1940-1944 range. American Photographer (January 1981) reported that an official from the High Altitude Observatory at Boulder visited the photo site and established a time of approximately 4:05 P.M. on October 31, 1941, based on relative position of moon and objects in the photo. In any case, sunlight and corresponding shadows are plainly visible on the grave stones. So one way to look at Moonrise is that it is a fake. The moon is fake, and the sky is even more fake. Despite being a single exposure (of sorts), it was made possible by deceitful filtration (Wratten No.15), selective underexposure, and other "analog trickery". Another way of looking at it is that it's a breathtaking image conveying stillness and solitude beneath the vast western sky.

Accepting that most shots with a full moon and any earthly content will require image manipulation, double exposure or skillful use of selective filtration, we can now examine the issue of correctly sized moons.

When I was kid my grandfather, whose father was in the civil war, told me that some folks thought the moon appeared as big as a wagon wheel, but he saw it as no larger than a silver dollar. Obviously this depends on how far away the dollar and wagon wheel are from your eye. So let's say, at arm's length, what familiar object is about the same size (occupies the same angular size) as the moon? This will surprise you.

All coins at arm's length are much larger than the moon. The moon, at about a quarter million miles distance is about as big as a pea at arms length. Both occupy about one half degree of angular space. With that in mind consider how big the moon will look in a photo, based on the type of lens used. The below table shows the angle of view of the short side of a 35mm film image, and of a 35mm digital body with a 1.6 magnification factor, such as the Canon 20D and Digital Rebel.

Lens length (mm)	angular size of short side (35mm) in degrees	angular size of short side (35mm -1.6 mag. factor)
17	56	46
50	27	17
100	14	8.5
200	7	4.29
300	4.57	2.86
400	3.50	2.14
500	2.75	1.72
600	2.29	1.43

When a full moon is directly overhead, it occupies around 0.52 degrees of the sky. This varies by about 12% depending on the moon's position in its elliptical orbit. On the horizon, a full moon occupies about 2% less space (or angle) because it is about 4000 miles farther away - half the diameter of the earth.

Near the horizon the moon is also slightly smaller because of the air-lens effect that is falsely claimed to be the reason the moon looks larger near the horizon. Also, this effect flattens the image of the moon, making its vertical angular size even smaller. For an excellent review of theories as to why we think the moon looks bigger on the horizon, see Donald Simanek's The Moon Illusion.

So regardless of position in the sky, the moon occupies about half a degree of angular space. Using the above table you can see that a shot of a full moon using a 300mm lens on a Canon 20D camera would result in a moon that spans a little over 1/6 of the short side of the image. Another way to view this is that you could fit 5 1/2 moons along the short side of a picture.

I have received several email messages asking about the "digital trickery" used in the above photo of the Golden Gate Bridge from Marin Headlands. I'm not sure why this shot evokes such suspicion. The moon is the correct size in the sky, and is in the correct position. It might be that its absolute size - actually, its size relative to the entire shot - is unusually large, or the fact that the photo is cropped very close (too close, by normal rules of composition) to the moon. The moon in this shot occupies 19% of the short side of the above shot of the Golden Gate Bridge, thus 5.23 stacked moons would span the width of the picture. This image is a slightly cropped shot from a Canon 10D with a 300mm f/4 lens. The city as shown in this image is technically underexposed, because two days before a full moon, when the moon is this high above the horizon, the sky is much lighter than it appears in this shot. The moon is also a bit blurry, partially because it was shot on the horizon, in the position where it appears in the photo. Haze and air density variations near the horizon produce a much fuzzier image than you'd get in the desert with the moon high in the sky. This exposure of the moon was also too slow to avoid motion blur. The moon moves across the sky about 1% of its diameter every second, so a sharp moon exposure, based on the traditional 1/1500-of-frame-width circle-of-confusion diameter, would require a shutter speed of 1/15 second or faster, which also pretty much rules out exposing a dusk or twilight city scene and the moon in a single exposure.

I find it amusing that the above Golden Gate moon is challenged as looking fake, meanwhile giant-moon posters of New York shot with wide angles lenses go unquestioned.

My friend Dan Heller (DanHeller.com also has many online tutorials, including one dedicated to moon photography) has published a number of popular shots with the Golden Gate Bridge and a full moon. His dramatic yet whimsical shot below includes an absolutely gigantic moon.

Dan makes no allusion to this being anything but a collage, and to clear up any doubt, has recently added information to his site about both exposures used in the image. He states that the bridge was shot with a 17 mm lens on an SLR camera. This can be easily verified, knowing that the bridge shot was taken from Battery Spencer, near the north tower of the bridge. Using a USGS map I determined the distance from Battery Spencer to the bridge's south tower to be right at 1.0 miles, or 5280 feet. The south tower rises 746 feet above the water. In Dan's photo the south tower spans two tenths of the height (short side) of a 35 mm film frame. From the above chart you can see that a 17 mm lens has a vertical (short side) view angle of 56 degrees. Therefore a 746 foot tower at a distance of 5280 feet should occupy aaround 0.23 of the height of the frame if the film plane was parallel to the south tower. But the south tower appears slightly below the cneter of the photo, indicating Dan was looking slightly upward, which would reduce the image height of the tower a bit. Close enough, then.

Now how about the moon? It occupies just under one quarter of the height of the frame, indicating it was shot with a 300 mm lens and cropped a bit. Or, perhaps it was shot with a shorter lens and enlarged in Photoshop, but that would have resulted in a lower quality moon. Interestingly, a moon properly exposed with Dan's 17mm lens would have occupied a tiny one percent of the height of the image - somewhere around the size of the red beacon lights along the suspension cables. You can use a similar approach to analyze many other moon/landscape photos.

Like the sun, the path of the moon changes from summer to winter. Many computer programs available on the internet can calculate the position of the moon at any time on any date. The motion of the moon is much more complex that that of the sun. Each full moon will rise and set in a markedly different location than that of the previous month. Also, many are surprised to find that unlike the sun, the moon stays closer to the horizon in the summer than in the winter. Stranger still, consider the full moon's path at two times of the year, one near the summer solstice and another near the winter solstice. In some years these paths will be fairly close together, while about ten years later they will be very far apart (The best explanation I have seen of this 19 year cycle is on Robert Pollock's Stones of Wonder site). In such years a summer full moon, in the northern hemisphere, can indeed be very close to the horizon, due south. So if we shrank Dan's big Golden Gate moon down to "proper" size, its position in the far southern sky above the bridge could be entirely legitimate. Even with the complexity of moon motion, a lot of moon shots could be judged fake (by more than just the exposure issue) based on position alone. The moon does not appear in the northern sky in Seattle. Ever.

Now we have established that nearly all moon-landscape photos are fake on at least one of several counts. Some are more fake than others, like those with overly large moons for the view angle of the photo, or with moons in strange locations. Now comes the issue of how a small 2-dimensional representation of large 3-dimensional space should best convey realism, or whether it should try to do so. You might say that at this point we enter into the realms of ethics and aesthetics. But at least now we know exactly where this point is.