And how come dimly lit worlds like Pluto look so bright in space photos?
There are a few questions that we get all the time at The Planetary Society. Look up at space at night from a dark location and you can see innumerable stars. Why, then, do photos of things in space not contain stars? How come the black skies of the Moon contain no stars in Chang’e photos?
The answer: The stars are there, they're just too faint to show up.
I can illustrate with an example from everyday life. I'm sure that everyone reading this article has made the mistake of shooting a photo of a loved one standing in front of a brightly lit window. In your photo, all you can see is a silhouette; your subject's face is a nearly featureless shadow. Their face still exists, of course! It's just not brightly lit enough to show up in the photo.
The same issues that can make your casual snapshots look bad affect space images, too. Let's talk about three things that affect what details you can see in any photo, whether it's of a star, a planet, or a person: the sensitivity of the camera, the time your camera had to collect light, and the dynamic range of your camera.
How Sensitive Is Your Camera?
How much light does your camera need to see by? Fancy cameras can adjust sensitivity by opening and closing the aperture that lets in the light. Human eyes do the same thing, automatically, all the time, by dilating and contracting their pupils. If you're a sighted person walking from a brightly lit to a dark outdoor area, you won't see stars in the sky either, at least not right away. As your eyes dial up their sensitivity by opening up your pupils, you slowly notice fainter and fainter stars.
Most space cameras actually can't adjust their aperture in this way. Instead, scientists predict the light levels that a camera will encounter through its mission, and design their instruments to have an aperture that's an appropriate size for the range of targets they expect to encounter. This can be a challenge if your spacecraft will encounter a wide range of target brightnesses, but you make your camera to work on the intended science targets and don't worry if it isn't ideal for any fun extras you may photograph along the way. OSIRIS-REx, whose MapCam was designed to study the colors of a very dark-toned asteroid, couldn’t look at Earth without getting overwhelmed by the brilliant light reflecting off of bright clouds, causing the artifacts you see in the top of this image.
NASA / GSFC / UA / Björn Jónsson
Earth from OSIRIS-REx MapCam
OSIRIS-REx flew past Earth on September 22, 2017 and took this photo shortly after. The Pacific Ocean covers nearly the entire visible globe. The Sun is nearly behind the spacecraft, and a bright area on the ocean near the center of the view is specular reflection from the watery surface. The image is a composite of three photos taken through infrared, green, and blue filters with exposure times of 1.5 milliseconds. The use of an infrared filter causes land that would appear green to appear red. "Icicles" at the top are caused by detector read-out register bleed-through, which occurs at the very short exposure times required for a close-up view of a bright planet.
But MapCam images of Bennu look fine, because that’s what MapCam was designed to photograph. (Note that in neither of these images can you see stars in the black space surrounding the worlds.)
NASA / GSFC / UA
Bennu from the south, 17 December 2018
This image was captured by OSIRIS-REx’s MapCam imager on 17 December 2018 as the spacecraft flew under Bennu’s south pole during the mission’s Preliminary Survey of the asteroid. The image has an exposure time of 9.3 milliseconds and was taken from a distance of about 12 kilometers while the spacecraft was traveling away from the asteroid.
How Long Does Your Camera Have to Collect Light?
Longer exposures collect more light -- helping to detect fainter things -- than shorter exposures. There's no real equivalent for exposure setting in human vision -- we notice more when we stare at something longer, but that's not quite the same thing. Photographs of the night sky that are full of stars are long exposures, often taking many minutes -- it takes that long for the camera to detect enough photons for a pretty view. Short exposures don't catch stars. The photo below was a minutes-long exposure. What looks like sunlight on the mountains is actually moonlight.
The Milky Way over Glacier Point, Yosemite National Park
The Milky Way from Glacier Point in Yosemite National Park. Moonlight still illuminates Half Dome and the high Sierras beyond. While the colors of the Milky Way are only visible in long exposure photographs, most of the detail is visible to the naked eye if you take the time to look.
The Apollo astronauts' photos were exposed for the brightly sunlit lunar surface and white space suits. These exposures were too short to detect stars in the sky.
Apollo 15 Commander David Scott at the ALSEP station, EVA 2
Taken during the second Apollo 15 Extravehicular Activity, on August 1, 1971. Scott is working at the site of the Apollo Lunar Surface Experiments Package (ALSEP).
Space cameras can permit a very wide range of exposure settings. New Horizons’ LORRI camera, for instance, can shoot images with exposures as short as 1 millisecond and as long as 30 seconds. They used the shortest exposure setting when they were flying past Jupiter, which is much closer to the Sun and much brighter than Pluto. They use the longest exposures for the faintest targets -- distant worlds in the Kuiper belt.
As an aside, this answers another common question we get about space images: how can cameras see to take pictures so far from the Sun, where the light is so comparatively dim? The answer: we send sensitive cameras and, if necessary, take long exposures. Voyager 2 at Neptune provides good examples of what happens when we send a camera that's not sensitive enough. Designed for Jupiter and Saturn, it had a tough time seeing in the relative dark at Neptune.
NASA / JPL
Voyager 2's best imaging of Neptune's moon Proteus
Voyager 2's cameras were designed for light levels at Jupiter and Saturn, so they struggled to gather enough light to see at Uranus and Neptune, especially when imaging dark targets like small moons. These are the best images that Voyager 2 took of Proteus. The upper left one, taken through a clear filter, is a 3.84-second exposure. The other three small ones, taken through color filters (clockwise from upper right: violet, green, and blue) required 15.36-second exposures, and two of them are blurred by spacecraft motion during that long time. The right one, taken through a clear filter, is a 1.92-second exposure; it's not blurred, but light levels are so low (and dark current of the camera is so relatively bright) that it's difficult to see detail.
What Is the Dynamic Range of Your Camera?
Is your camera capable of seeing both dimly-lit and well-lit things in the same image? Or does its light-collecting capacity get quickly overwhelmed by brighter things before it's had time to detect any light from dimmer things? Here is where our eyes generally do much better than our cameras. When I see a friend sitting in front of a window, I can see their face just fine because my eyes are capable of discerning detail in both shadow and sunlight. This is partly because my eyes aren’t still when I look at a scene. My eyes constantly dart about, looking out the window, looking indoors, looking at my friend’s face, each time adjusting focus and aperture. My brain builds up a composite of all this information, making the view in my mind’s eye more detailed than any instantaneous view from my physical eye. Then I take out my camera and take a picture, and it looks terrible.
But wait -- modern digital cameras have a trick that mimics what the human eye and brain do. With my phone camera I can turn on a feature called “HDR,” which stands for high dynamic range. When I take a photo in HDR mode, the phone actually takes two photos (one longer-exposure, one shorter) and merges the best-exposed parts of both images, to show me details both outside the window and in my friend’s features.
Space cameras typically have higher dynamic range than consumer cameras, so are able to record relatively faint and bright things in the same image. It can be hard to appreciate just how much detail there is in the shadows of space images, because our everyday digital displays are mostly not capable of such high dynamic range. But you can play with brightness and contrast in space images to reveal hidden details in the shadows. Check out what’s visible in two different contrast stretches of a Rosetta OSIRIS image of comet Churyumov-Gerasimenko. It's the same photo, made from the same data, I've just told the computer to display low pixel values with higher brightnesses.
ESA / Rosetta / DLR / MPS for OSIRIS Team MPS / UPD / LAM / IAA / SSO / INTA / UPM / DASP / IDA / Emily Lakdawalla
Before & after: Fractures in shadowed Churyumov-Gerasimenko cliffs
Enhancing the brightness of OSIRIS images can reveal details in shadows only indirectly lit. Here, some cliffs on the head of the comet are revealed to have at least two criscrossing sets of fractures.
Example: New Horizons
Let's explore how exposure settings and other factors make stars visible sometimes and not at other times. Let's travel together aboard New Horizons as it uses its LORRI camera to take pictures of targets with varying intrinsic brightnesses, catching some stars along the way. As I mentioned earlier, LORRI has a fixed aperture and high dynamic range and can use a wide range of exposure settings. One of the coolest targets at Jupiter is the volcanic moon Io. Here’s a LORRI picture of Io taken as New Horizons approached the Jupiter system. The photo used a 4-millisecond exposure. Io is well-exposed, and no stars are visible.
NASA / JHUAPL / SwRI
Io from New Horizons (exposure time: 4 milliseconds)
New Horizons took this photo of Io with its LORRI camera from a distance of 2.865 million kilometers on 26 February 2007.
A few seconds later, LORRI took another photo of Io, with a much longer exposure: 75 milliseconds. Io is severely overexposed, and the effects of that overexposure are causing something called “readout smear,” the streaky effect across the image. Why would they do that? Check out the edge of Io’s disk. You can see at least three volcanic plumes erupting off Io’s surface. At this exposure setting, there are rich details visible in the plumes. Still, even with an exposure almost 20 times longer than the above Io image, I’m not convinced that any stars are visible. (I'm not sure what the one dot to the right of Io is, but it's the only one in the photo, and is very bright; I think it's probably a cosmic ray hit.)
NASA / JHUAPL / SwRI
Io from New Horizons (exposure time: 75 milliseconds)
A long-exposure image of Io taken on 26 February 2007 during New Horizons' Jupiter flyby. Sunlight on the dayside of Io has saturated the New Horizons LORRI camera's detector, causing the streaking across the photo. But the long exposure reveals lovely structures in the umbrella-shaped plume of the actively erupting volcano Tvashtar. Two or more other plumes are visible on the night side of Io.
Here’s a different set of Io images. By this time, 2 days later, Io had moved around Jupiter and into Jupiter’s shadow. No sunlight is hitting Io’s surface. You’re seeing Io glowing away in the dark, lit by its hot volcanic plumes. (You’re also seeing a lot of camera artifacts caused by stray light -- Io was in eclipse, but New Horizons was not; Jupiter’s light is bouncing into the camera, causing splashes of light and a bright background. There is also a ton of fuzz caused by energetic particles hitting the detector -- Jupiter has a lot of that stuff flying around.) Think for a minute about how cool it is that we can see the hot glow of Io's volcanoes in the utter dark of an eclipse! It took a much longer exposure, nearly 8 seconds, to make Io’s self-light from its volcanoes visible to LORRI. Finally we have a long enough exposure to see stars in the background. In fact, I’ve used those stars to align the images; in this animation, stars hold still while Io moves around and New Horizons adjusts its pointing.
NASA / JHUAPL / SwRI / Emily Lakdawalla
Io in eclipse from New Horizons
As New Horizons passed by Jupiter, it saw Io move into Jupiter's shadow. With the Sun blocked from Io's surface, New Horizons could see the light emitted by Io's volcanoes. Each image in this 28-frame animation required an 8-second-long exposure to make the faint glow of the hot volcanoes visible. The long exposures also make background stars visible. There are a number of artifacts in these images. Large streaks across the photo and the generally bright background result from stray light entering LORRI's barrel -- sunlit Jupiter is not far outside LORRI's field of view, and its brilliant clouds are bouncing light into the camera optics. There is also a lot of "snow" from energetic particles striking the detector.
Here: to help you, I’ve picked out some of the stars. Concentrate, and you may see some more; I’m pretty sure there are lots of fainter ones in the upper part of the animation.
NASA / JHUAPL / SwRI / Emily Lakdawalla
Io in eclipse from New Horizons (annotated)
As New Horizons passed by Jupiter, it saw Io move into Jupiter's shadow. With the Sun blocked from Io's surface, New Horizons could see the light emitted by Io's volcanoes. Each image in this 28-frame animation required an 8-second-long exposure to make the faint glow of the hot volcanoes visible. The long exposures also make background stars visible. In this version of the animation, brighter stars are marked with horizontal white lines. There are a number of artifacts in these images. Large streaks across the photo and the generally bright background result from stray light entering LORRI's barrel -- sunlit Jupiter is not far outside LORRI's field of view, and its brilliant clouds are bouncing light into the camera optics. There is also a lot of "snow" from energetic particles striking the detector.
(At the risk of making things too complicated, I'll mention that actually, New Horizons used a second trick, besides long exposure, to make faint things more visible in these eclipse images. They binned the data, averaging together 4-by-4 groups of pixels, and so returning the images only 256 pixels square instead of 1024. By binning the data, they get 16 times fewer pixels, but they make the camera that much more sensitive, less affected by random noise. Binning the data also prevents images from blurring from spacecraft motion during long exposures.)
As New Horizons approached Pluto, stars were still visible in their relatively long exposures. You can still see stars in these approach animations; below, amateur Matthew Earl used the stars to align the images and reveal the funky dumbbell rotation of Pluto and Charon.
NASA / JHUAPL / SwRI / Matthew Earl
New Horizons' optical navigation images of Pluto and Charon aligned on background stars