Astronomy with a cellphone — how low (faint) can you go?
Preface: An unfortunate aspect of astronomy as a hobby is the impact of weather. It doesn’t take much in the way of clouds to render going outside pointless. The week centered around full moon creates further limits. Here in the Boston area we‘ve maybe averaged 4 nights/month that are usable this year. Some of these posts sit around waiting for pictures — and then forgetting about finishing the writing for many months.
I’ve tried to avoid using the word dim as it implies a property of an object that would be the opposite of bright. All astronomical objects that we can see directly are either emitting or reflecting light, so they are all bright, just some more so than others. It’s like describing temperature, in an absolute sense there is no such thing as cold, just objects that are not as hot as other objects.
Adding a lens to the cellphone provides magnification but the primary motivation is to capture more photons, which means we can take pictures of fainter objects. A lens can also reduce the impact of light pollution on your pictures and make those fainter objects stand out better, though the best improvement for that problem (assuming you can’t arrange for a power blackout across your state) is with multiple exposures and post processing.
Most of the astronomy pictures you see are of nebulous or extended objects like galaxies where due to their great distance the individual light of those billions of stars merge in to one larger fuzzy object. With a few exceptions using your cellphone or cellphone plus the 60mm lens used here won’t capture these extended objects, though there’s some exceptions.
One way to gauge the light sensitivity of your cellphone is to take a picture of a section of the sky and compare it with a star chart that provides stellar magnitudes. The first hurdle to overcome is matching your picture against a star chart, the technical term for this when it applies to a software solution is called plate solving. Name your pictures by the constellation(s) they feature to help you remember what they are.
When looking straight up you are looking though as little atmosphere as possible. The more air you look though as you look away from straight overhead the fainter the object will appear. Due to air movement the image may be more unsteady and the stars appear blob like.
When discussing how bright something is it’s helpful to understand the way astronomers describe brightness. That scale is logarithmic; you did pay attention to logarithms in your high school math class, right? Logarithms are a fundamental mathematical concept and are found in equations that describe everything from rush hour traffic to the history of the entire universe.
Something that is one stellar magnitude less than something else means it’s 2.5 times brighter, i.e. very dim things have large magnitude numbers, very bright things have very small or even negative magnitude numbers. A three magnitude difference is 2.5 x 2.5 x 2.5 =15.6 times different. For examples and more on the topic check out Wikipedia.
The picture below was taken with the Pixel 3a in its default astrophotography mode and the 60mm Sirui lens. Taken in mid June from Cape Cod in Massachusetts, it’s centered on Albireo which was about 45 degrees above the horizon at the time the picture was taken.
With the picture above it’s tough to decide where the constellations are. Most people living in light polluted skies so normally don’t see the fainter stars and it can take some effort to identify things.
Another effect for photos is that for brighter stars the pixels are all solid white, the dynamic range of the captured image isn’t high enough to capture bright stars as being brighter than others. Instead what happens is the bright stars appear bigger in a photo, which is not what your eyes see looking up at the stars.
If you’re having trouble figuring out what you’re looking at here’s the picture with annotations.
Your computer display can’t display a wide range of brightness values; it’s probably limited to around 5 (astronomical) magnitudes and these pictures cover around 8 magnitudes of brightness, or 15 times the range of your display. Desktop planetarium software like Stellarium resort to size to convey brightness.
It is possible to adjust the settings to show more stars in a large field view like above, but it becomes confusing.
When zooming in on a region Stellarium (and other similar applications that have a reasonable catalog of stars) will show the fainter stars. For example a zoom in to the constellation Lyra from the phone’s picture is shown below, the second picture is a screen capture from Stellarium.
Vega appears larger due to diffraction and imperfections from the optical path. The blue flare around Vega is from chromatic aberration, and being near the edge of the lens field there’s some coma too. Astrophotography exposes optical flaws that go unnoticed using your phone to take vacation pictures. We’re not aiming for perfection with the phone setup — we just want to see what it can do and understand the limits.
The image from the phone can be compared with Stellarium (or other planetarium software) to determine the faintest stars that the phone can capture in its default single 60 second exposure. In the case of the Pixel 3a the phone captures around ten 6 second exposures and processes those to create the final image.
It’s much easier to do this comparison placing images side by side.
When annotating star charts with magnitudes the convention is to not use a decimal point as it could be mistaken for star and vice versa. After the star Sulafat (mag 3.3, written as 33) the next brightest star in the image is λ Lyrae (Lambda Lyrae, HIP93279) at mag 4.9 (written as 49). λ Lyrae is a K2 type star so has a more orange color compared to our own sun’s (G2) yellow color. The picture captures that color and Stellarium simulates how it would appear through a telescope.
The dimmest star marked is mag 9.3 star near λ Lyrae, though it’s dimmed to about mag 9.4 since it wasn’t overhead at the time of the picture. The star is not named in a catalog but has J2000 coordinates 19h00m43.11s /+32°20'56.9".
Looking at the two photos side by side we can see another factor that would limit the faintest star we can capture: the picture’s background is a dark green compared to the black background of Stellarium.
That greenish background could be from several items, but most is probably the background light pollution from where the picture was taken. There is natural sky glow as well as the overall light of billions of background stars in this part of the Milkway disk. The green color may be because the camera chip in cellphones use what’s called a Bayer filter to create a color image so are naturally more sensitive to green.
Post processing and/or combining more images could reduce the background brightness and make the faint stars appear brighter.
Our reference picture also illustrates a side effect of the Pixel 3a’s processing: it tends to exaggerate color in some cases. Go back to the original picture and look for Brocci’s cluster (Collinder 399), also known as the coat hanger. Brocci’s cluster is not an actual open cluster of gravitationally bound stars. The stars we see in it span a distance of over 1000 light years. Instead it’s a random orientation of stars along a line of site that our pattern matching brains stuff in to an imagined pattern. Not a thing known to the ancient Greeks, since coat hangers as we know them now were a 19 century invention. All hail the magnificent heavenly coat hanger!
Zooming in further on the photo, the stars have a colorful presentation, particularly when compared to Stellarium screen capture in the next picture.
The magnitude 6.3 star looks strongly greenish. It’s actual color should look about the same as the magnitude 6.6 star to the lower left of it. In real life stars are almost never perceived as green by the human eye as hot enough object appears white to the human eye. We refer to “white hot” coals, not green hot.
Without knowing the algorithms used by Google’s photo application it’s difficult to say exactly why that star is green. This tendency for wrong colors has been seen in other astrophotography pictures captured with the Pixel 3a.
Looking at the zoomed in view we can see that the phone captured stars down to the magnitude 9.0 range — closer to 9.2 if we allow for atmospheric extinction from the stars in this part of the photo. The puts over 100,000 stars within reach of the cellphone’s picture capabilities (based on this estimate and this one). Not bad for a 3 year old cellphone.