 |
Path: physics insights >
misc >The Lack of Stars in the Apollo 11 Lunar Landscapes |
First, however, we need to determine what the assertion means. It's presented as a statement of fact, and as such it should be testable. However, it contains an undefined term: "should"; statements regarding how things "should" be are intrinsically subjective and hence not testable. So, before we can test the assertion, we need to rephrase it in such a way that it's both objective and more specific. So, the assertion we will actually test is the following:
"Using knowledge obtained from experiments performed on Earth, extrapolated to conditions as we believe they exist on the Moon, we can conclude that a properly exposed photograph of a lunar landscape in bright sun would also show images of the stars in the lunar sky."On the remainder of this page we will describe some simple experiments which were performed, along with some simple extrapolations from Earth conditions to lunar conditions, which will lead us to a firm conclusion regarding the possibility of showing star images in full-sun lunar landscape pictures. As always, we prefer to do our own experiments rather than "appealing to authority", so the data will come entirely from photographs taken locally, with some additional data obtained from past experience with manual exposure cameras.
And, in fact, after carrying out the experiments and the analysis, the conclusion we have come to is that the above assertion is false -- the stars would not be visible in a properly exposed lunar landscape picture.
Now let us start by considering the terrestrial data, and then we will discuss the extrapolations to lunar conditions.
Exposure Latitude
First, some comments on the film used on Apollo 11, in the light of knowledge obtained in Earth-side photography.NASA says that the film used in the still camera was Ektachrome 160. It's a film I used a fair amount, back around 1970 when it still went under the name of "High Speed Ektachrome". The film I used was the "amateur" version, while the Apollo crew was using a special version, no doubt based on the "professional" film of the same name. However, the basic properties of the film were the same for the amateur and professional versions.
That film was somewhat grainy and rather unforgiving: its exposure latitude was narrow. An exposure which was half a stop under might appear more saturated (which was considered good) but might also appear a bit muddy, and rather grainy. A full stop under was definitely looking muddy and very grainy, and anything farther off than that was pretty much useless. Similarly, slightly overexposed frames lost their highlights and looked washed-out; significantly over-exposed frames were useless. The flip side of the lack of latitude was high contrast: Dim objects looked darker than they were, and bright highlights would "burn out" and go white. Very dark objects would disappear into a uniform dark sea of "d-max", losing all detail. This last point is worth keeping in mind when we ask about dim stars showing up in a dark sky in a frame exposed for the brightly lit foreground objects.
Why did NASA choose this film? I don't know for sure but I can certainly guess. The camera was not auto-focus -- no cameras were in 1969 -- and the astronauts may have found it awkward to carefully focus each shot. So, they would probably have wanted the option of closing down the lens to a small aperture for maximum depth of field, setting the focus to the "par-focal" point, and hoping for the best; that would work for most "middle distance" shots, with no fiddling with the focus between shots. But they also would have wanted to keep the shutter speed high enough to avoid camera shake, as the camera was generally being used off-tripod. That spells a need for a fast film. They also wanted color. There was a very limited selection of fast color films available in the late 1960's, and high speed Ektachrome was about the best of the bunch. There's a good chance they also specifically wanted a slide film, as the colors were generally truer and brighter and the results were felt by many to reproduce better than shots made on print film. (National Geographic, for example, required all photos submitted to the magazine to be slides back in those days. In fact every picture in that 'zine was marked in fine print in the margin, indicating whether it was Kodachrome or Ektachrome; as far as National Geographic was concerned, there apparently was no third choice.)
High speed color print film such as Kodacolor Gold was still a decade away, and high-speed high-latitude fine-grain "super films" such as Kodak Max 800 were an impossible dream back in 1969. The large grain and the high contrast were a direct consequence of the high speed of the film; using 1969 technology, enhancing film speed was done, in large part, by increasing the grain size, which increases both apparent grain and contrast; the latter inevitably reduces latitude. With a large-format camera such as the Hasselblad the grain was not a big concern, but the larger format does nothing to help the latitude, or control the contrast.
The Camera Used in the Tests
The test shots on this page were taken with a Canon PowerShot A530 electronic camera, operated in manual mode. Compared to a modern-day superfilm like Kodak Max, the CCD in the camera has mediocre exposure latitude, and isn't great at showing detail in dark shadows and bright highlights. However, it stacks up pretty well against most of the slide films I have used. In short, its performance is a reasonable match to the old Ektachrome 160 used on Apollo 11.For most shots the camera sensitivity was set to ISO 100, which provides "normal" behavior, similar to that provided by analog "wet-darkroom" film. For the last couple of shots made by starlight, the sensitivity was increased to ISO 400 or ISO 800. Unlike the old analog films, this doesn't seem to have much effect on the contrast, latitude, or grain, but it does dramatically increase the "noise level" of the images; consequently the high-ISO shots provide useful data but lousy aesthetics.
Finally, these are all 24 bit images (so-called "32 bit"), which actually only have eight bits of grayscale (brightness) information in them. This is not ideal; however, it is identical with the available digital scans of the NASA images, and it's likely to be far better than any published print made from the NASA slides. With respect to the NASA images, you could only improve on the information content of a 24-bit scan if you obtained access to the actual slides, or to a large-format analog print made from an internegative made directly from the NASA slides.
Images shown on this page have been scaled down but, except as noted in the captions, are otherwise unmodified.
Daylight Exposures
We define a "standard daylight exposure" as an exposure made at F/16, with the shutterspeed set to the inverse of the film ISO rating. Thus, at ISO 100, the standard daylight exposure would be F/16 at 1/100 second.On this web page, we'll define the "Exposure Value" or "EV" as measuring how much less exposure was given than the standard daylight exposure value, measured in "stops", where each "stop" is a factor of 2. If the ISO rating for the CCD or film used in the exposure is S (for "speed"), the shutterspeed is T ("time"), and the F stop is A ("aperture"), then the formula for this is
(1)
For ordinary terrestrial landscapes under bright or hazy sun, the correct exposure is EV 0. For other conditions, the expected correct exposures are as follows:
Table 1:
| Bright sun on sand or snow (highly reflective surroundings) | EV +1 |
| Bright or hazy sun, distinct shadows; ordinary surroundings (not highly reflective) | EV 0 |
| Hazy sun, soft shadows | EV -1 |
| "Cloudy bright" -- light overcast, no shadows | EV -2 |
| Heavy overcast, or "open shade" (subject lit by sky and reflected light, but no direct sun) | EV -3 |
These are the values I used for years before I had a camera with a meter. When writing this page, I verified it against information from Kodak's website. As we will observe later, there are a couple of interesting things about this table (which are amply confirmed by personal experience):
- The only compensation for the reflectivity of the surroundings which is needed is in the case where the surroundings are bright white. With bright sun on sand or snow, we need to reduce the exposure by about a stop (a factor of 2). But dark, non-reflective surroundings require no such compensation; in other words, under normal circumstances (surroundings not bright white) the bulk of the illumination is coming directly from the Sun or from the sky.
- The value for open shade is EV -3. That is, a subject lit by sky light and reflected light from the surroundings typically receives about one eighth the light of a subject lit directly by the sun.
This implies that the bulk of the illumination of a subject in direct sunlight is coming directly from the Sun -- not from diffused light from the sky, and not from reflected light from the surroundings.
EV +1: ISO 100, f/5.6, 1/1600 sec | EV 0: ISO 100, f/5.6, 1/800 sec |
EV -1: ISO 100, f/5.6, 1/400 sec | EV -2: ISO 100, f/5.6, 1/200 sec |
EV -3: ISO 100, f/5.6, 1/100 sec | EV -4: ISO 100, f/5.6, 1/50 sec |
The daylight test shots came out much as we would expect:
- EV +1
-- This is the "correct" exposure for bright sun on sand or snow.
As we can see there's a lot of snow lying around in these
pictures, so even though there was some thin cloud cover, we would
expect this to be the "correct" exposure. The sunlit bricks in
the upper story look dark brick-color, the shutters look black, the
trees and bushes look about right. The snow looks a little dark
to me, but it has good detail.
- EV 0
-- This would be the "correct" exposure if the snow weren't there, and
it actually looks pretty good. The bricks look good, and the
bushes look good. The black shutters might be just a touch pale,
though, and the snow is definitely "burned out".
I'd put the "exactly correct" exposure somewhere between these first two, around EV +0.5 or so. Like Ektachrome 160 film, the CCD used in this camera is rather unforgiving of exposure errors, as we can see in this sequence of shots. - EV -1 --
Correct for hazy sun with distinct shadows. We had somewhat hazy
sun that day but the snow more than made up for it; this is clearly
overexposed. However, the porch, which is shaded from the sun,
still looks underexposed to me.
If we look at the area just under the overhang on the second story, we get an idea of how much light is being provided by reflection from the snow. The brightness of that area appears to be between the brightness of the sunlit bricks in the EV+1 and EV 0 shots. This implies that the snow is providing roughly as much light as the direct sunlight -- which is why we need to go to EV+1 for bright sun on snow, of course. - EV -2
-- Correct for "cloudy bright", this is clearly 'way over for this
scene. However, if we look at the porch, it looks pretty good,
and the pine tree in the background toward the right looks like its
inner branches, which are shaded from the direct sun, are looking
pretty good.
The porch is "open shade", and would normally be around EV -3. But with all the snow brightening the scene, it looks like EV -2 is about right for this patch of "open shade". - EV -3
-- Correct for "open shade", this is heavily overexposed for this
scene. Even the porch is looking overexposed. If we look at
the mailbox, to the left of the door, it looks about right -- but it's
shaded from a great deal of the "snow light".
- EV -4 -- One stop below "open shade". For this scene it is very overexposed, with detail appearing only in a few of the shaded areas of the picture. Most of the house and bushes have vanished; they're entirely burned out.
From all this, we can conclude that the exposure table is pretty accurate for this camera. The snow has apparently pushed everything "up" the scale by about 1 EV -- but even without compensating for that, exposures made according to the table would be close to correct.
We also can conclude that any scene photographed in bright sun which looks "mostly correct" must have been photographed at about EV 0. Certainly a brightly sunlit scene which was photographed at EV -3 or lower would look very washed out, and one photographed at EV -5 or lower would be nearly blank white. The Apollo photos taken on the moon look "about right" so we might conclude that they must have been taken at about EV 0 (but we'll have more to say about this later when we discuss atmospheric attenuation).
Starlight Shots
On Earth, of course, the blue sky "drowns out" the stars. So, we need to wait for darkness to find out how bright the starry sky is, and find out if it would show up in a daylight shot if only the blue sky weren't swamping the starlight. This next sequence of photos was taken at about 11 PM, under similar conditions to the earlier daylight shots: The landscape was snow-covered and there were some thin clouds, but it was mostly clear. This location is also near a city so there is some sky glow.Because there was -- and always is, on Earth -- some haze in the air, and because there is some light pollution, we're not just interested in determining whether there are recognizable images of stars in the following images. Rather, we want to know when there is anything visible in the images -- at what exposure are the frames no longer blank black? That is the exposure value at which we might reasonably expect to record images of stars, if the sky between the stars were pure black.
The table of photos follows, and commentary on them is given after. As with all pictures on this page, except as noted, the images have been scaled but are otherwise unmodified.
More photos were taken, but it seemed pointless to include a large number of blank frames on this page so I've just shown the highlights, so to speak.
We'll now discuss the starlight photos in some detail.
- EV 0: This is exposed at a "normal value" for a daylight shot in full sun. The result is what is technically known as a "blank frame".
There are no pixels brighter than about level 4 (out of 256
brightness levels), as we confirmed by converting to black and white
with a 'threshold' of 4. Nothing registered -- no stars, no sky
glow, no clouds. It appears just as though it had never been
exposed.
This is the velvet black sky we see in the Apollo photos. - EV -7.7: This is also a "blank frame". Threshold of 4 again showed no white pixels.
Our earlier tests suggest that, if the Apollo pictures had been exposed at EV -7.7, they would have been totally washed out.
A photograph taken in full daylight at this exposure would be overexposed by a factor of 200, yet this night shot shows nothing. - EV -9.7:
This shows nothing brighter than about level 25 (out of 256).
Stars are not visible, no matter how it's enhanced; we're seeing
a little noise and not much else.
Since this photo shows some noise, and is not completely black, we can say it is brighter than the Apollo skies, which appear almost uniformly black. So, this suggests the Apollo photographs were taken at an exposure value greater than EV -9.7 (which should come as no surprise, as they appear to have been exposed at EV 0).
A photograph taken in full daylight at this exposure would be overexposed by a factor of 800, yet no stars, clouds, tree silhouettes, or other items are visible. - EV -11.7:
Very little shows in this image. Almost everything is at a
brightness of 28 or less out of 256 levels). The Pleiades are
down around a brightness of 14 or dimmer. With histogram
equalization, some detail becomes visible, as shown in the "enhanced" version (equalization done via PhotoImpact program).
A photograph taken in full sunlight at this exposure would be overexposed by a factor of 3,300. - EV -15:
There's something in this image, but not much -- we're still not
looking at a night sky photograph as we typically think of it.
With vigorous enhancement, we can see the
Pleiades; thus, the stars produced recognizable, but very dim, images.
(The enhanced version was produced using GThumb's "enhance"
feature.)
A photograph taken in full sunlight at this exposure would be overexposed by a factor of 30,000. - EV -17.8:
We have good images of stars, the sky glow (light pollution) is
clearly visible (the sky is not black, but is dark gray), and
clouds are visible. We have arrived at a reasonable night sky
picture. If not for the sky glow, this would look like what you
might have pictured when you thought about the sky on the Moon.
However, a sunlit landscape at this exposure would be overexposed by a factor of 227,000. It would be blank white, totally washed out. The Apollo landscapes were obviously not shot at this exposure. - EV -18.8: An even better exposure; this is a good setting to use for photographing stars at my location.
A sunlit landscape shot at this exposure would be overexposed by a factor of 454,000.
Combined with the results from the previous section, which indicate that properly exposed landscapes such as the Apollo images must be shot at about EV 0, this leads us to conclude that the skies in Apollo pictures should be velvety black, without stars, just as they appear. However, we're not done yet, as we'll see in the next section.
Lunar Conditions versus Terrestrial Conditions
The moon has essentially no atmosphere, which means the light from the stars is neither absorbed nor diffused. Furthermore, the reflectivity of the lunar surface is terribly low (about 7%), as a result of which the illumination of a sunlit scene on the Moon may be lower than expected. In the following sections we'll consider these issues in more detail.Attenuation by the Atmosphere
The light from the stars must pass through about 200 miles of atmosphere to get to the Earth's surface. That surely attenuates the view we get; the stars are surely brighter on the Moon. However, this effect can be ignored, because it is exactly compensated by the attenuation of the Sun's light, which must also pass through that same 200 miles of air.In case this isn't clear, I'll discuss it in a little more detail. At the Earth's surface, the Sun's light will be attenuated by exactly the same factor the starlight will be reduced by. In consequence, a properly exposed lunar landscape will need to be shot at a higher EV than a properly exposed Earth landscape -- and the difference will be exactly the same as the difference in how easily imaged the stars will be. If we suppose the starlight is attenuated by, for example, 8 EV on Earth, then we must suppose the sunlight is also attenuated by 8 EV also; lunar landscapes will therefore be shot at EV +8 rather than EV 0, and the stars, which would be visible in a lunar exposure at EV 0, will actually be invisible in a shot which is correctly exposed under the unattenuated sunlight of a lunar landscape.
Consequently attenuation of starlight by the atmosphere is not an issue.
Light Scattering by the Atmosphere
Scattering of sunlight -- which is what makes the sky blue -- might appear to be more of a problem. The reason is that all the scattered light from the Sun which shines down from the sky eventually makes its way to the subject, but scattered light from the stars just contributes to sky glow; if the sky glow is too dim to record an image on film the scattered starlight will essentially just be lost. This would tend to reduce the apparent brightness of the stars at Earth's surface while leaving the effective brightness of the Sun unchanged. (This was touched on in point 2, above.)However, we can estimate how much scattering we're actually dealing with just by looking at the proper exposure values for conditions of less than full sun (see table 1). A subject which is lit by scattered light from the sky, but which is shaded from directly sunlight, is said to be in open shade.
The proper exposure for open shade is about EV -3, which is 8 times the exposure given to a subject in full sun. In other words, a subject in sunlight receives about eight times as much light directly from the Sun as it receives from light shining on it from other parts of the sky. From that, we can conclude that only about 1/8 of the light is scattered and diffused; the other 7/8 travels directly to the subject.
That in turns leads us to conclude that, while this effect is real, the best estimate we can put on its magnitude is that it would reduce the apparent intensity of the stars relative to the Sun by about 1/8, or about 12%, at the Earth's surface. This is far less than 1 EV, and certainly not nearly large enough to affect our conclusions.
Lunar Surface Reflectivity
The lunar surface is very dark. This could conceivably result in a need to expose lunar landscape photos at some lower exposure value than EV 0, which would in turn make the stars more visible. (This was touched on in point 1, above.)We can estimate the degree to which this might affect the results by looking at the exposure chart for terrestrial landscape photography (table 1). When the surrounding surface is bright white, such as sand or snow, we need to compensate for it, and expose at EV +1. However, there is no compensation for dark surroundings. In any locale in which the surroundings are not bright white, when exposing a subject in bright sun, we can safely ignore the contribution of light reflected onto the subject by the landscape.
This point is brought home by looking again at the exposure value for "open shade". A subject in open shade is typically light by a large area of sky and light reflected by the landscape. Proper exposure in "open shade" is typically EV -3, which is eight times the exposure given to a subject in bright sun. In other words, the total contribution to subject illumination, from sky light and light reflected from the landscape, is typically about one eighth of the illumination provided by direct sunlight. In consequence, while some compensation might be needed when photographing a subject on an exceptionally dark landscape, the resulting change in exposure would be substantially less than 1 EV.
This leads us to conclude that, while the effect is no doubt real, it is not nearly large enough to affect our conclusions.
Sky Glow
Sky glow -- light pollution, and light scattered from stars and other natural sources -- certainly makes the stars more difficult to see on Earth. Does the lack of sky glow on the Moon mean the stars should show up more easily in photographs?No, it doesn't -- sky glow doesn't make the images of the stars dimmer, it just makes the spaces in between the stars brighter. In fact, light pollution actually makes the images of the stars slightly brighter, while simultaneously "drowning them out" with bright sky between the stars.
In our starlight shots, down to EV -7.7 we saw nothing in the frame. The stars -- and the sky between them -- simply didn't register at all. At EV -9.7 we started seeing something (but no identifiable star images). Thus, there is a "floor" of about EV -8 to EV -9; exposures taken at higher EV values show a pure black sky. Sky glow actually raises this "floor" value, by adding to the brightness of the night sky. On the Moon, with no sky glow, we would expect to be able to take exposures at somewhat lower exposure values while still seeing nothing but a pure black sky!
Conclusions
From the foregoing test shots and the accompanying reasoning, we are led to the conclusion that, in a properly exposed photograph taken of a subject in bright sun on the lunar surface, the stars would not be visible. The sky would be expected to appear as an undifferentiated black area.That is, indeed, exactly what is seen in the Apollo lunar landscape photographs.
Some Additional Considerations
When the assertion is made that the stars "should have been visible" in the Apollo lunar landscape photographs, it's often accompanied by some other assertions which are even less well grounded. We'll mention just a few of these in passing.First, though, we should point out that there have been many, many photographs taken in low Earth orbit. Using a camera a few hundred miles above the Earth is clearly not a problem.
In low Earth orbit, nearly the entire atmosphere, and its shielding abilities, are below the astronauts. They are entirely exposed to space -- except for one thing: The Earth's magnetic field, which extends far out beyond low earth orbit. The magnetic field traps charged particles, and hence provides shielding from certain kinds of radiation in LEO. But it doesn't have any effect on electromagnetic radiation.
Now, let's consider just a few other claims one sometimes hears.
- "X-rays
from the sun would have ruined the film if one took a camera to the
Moon." I heard this claim on a YouTube video, or I would not
have believed anyone could ever seriously say such a thing. It's wrong
on multiple counts.
First, the Sun emits very weakly in the X-ray band -- it's certainly not going to be fogging film inside a camera. (The Sun is not an X-ray star.)
Second, and more to the point, X rays are not affected by the Earth's magnetic field. So, if photography on the Moon were impossible due to the X radiation from the Sun, then photography in low earth orbit would be equally impossible! - "Intense
ultraviolet from the sun would have washed out every picture."
This claim is absurd; it, also, comes from a YouTube video
"debunking" the Apollo flights.
First, to believe this you either need to know nothing about ultraviolet light and photography, or you need to assume that the people at NASA, Hasselblad, and Kodak are all idiots who don't know the Sun produces ultraviolet light. UV is trivial to filter out, and UV filters were common accessories back in 1970. One for a 50 mm (wide angle) Hasselblad lens would probably have cost about ten dollars. Could they really simply have forgotten to purchase a skylight filter when they bought that custom made camera? Or, rather, does it make sense to assume that they must have forgotten the UV filter, and then try to conclude from that, that the pictures couldn't possibly have turned out?
But again, and more to the point, UV is not affected by the Earth's magnetic field, and it's every bit as intense in low Earth orbit as it is on the Moon. So, if photography on the Moon is impossible due to the intense ultraviolet light, then photograph in low Earth orbit must be impossible, too. And we have an awful lot of photographs taken in LEO which show that isn't so. - "Temperature extremes would have made the camera fail."
They would? Prove it, please. The camera was
specially engineered just for the moon flight; it seems pretty silly to
simply make a blanket statement that the engineering problems are impossible to solve, without providing any evidence or reasoning to back it up.
But in any case, we again can look to LEO flights. The same exact temperature extremes affect every satellite in orbit around the Earth -- there is nothing special about the lunar environment in this regard. If it's impossible to make a camera that will work in such extreme conditions, then all the cameras in orbit around the Earth must be impossible, too. (And that includes Hubble, of course.)
And this is enough for now.
| |
Page created on 02/11/2008. Typos corrected on 02/12/08.