Making an Astrophotograph

Making an astrophotograph starts with equipment. For this image, I used a stock Canon 6D that I have had for years and a Canon 50mm f1.8 lens that I bought for $100. I put the camera on a sturdy tripod that I have also had for years. One piece of equipment that I bought specifically for astrophotography that I had no need for in daylight photography is an intervalometer which is a tool that allows you to take many photos remotely without have to press the shutter release for each photo. Whatever camera is used, it needs to have the ability to shoot images in a format that doesn’t throw away data for the sake of compression. Formats such as jpg make files smaller by throwing away data it has deemed unimportant. Typically, this means that pixels with light levels above or below a certain value are thrown away. This is called clipping. Obviously, in low light photography, data at the low end of the light level range is important and we don’t want to throw it away. So I shoot my images in raw format which means the files are very large with no data thrown away.

The first step in taking a photo is focusing. Focus in astrophotography is surprisingly challenging at first. To get sharp focus, I set the ISO of the camera to as high as it will go. I find a bright star and put it in the center of the viewfinder. I switch the camera to live view and zoom in on the star as far as the camera will allow in live view. For my camera that is a 10x zoom. I focus by making the star as small as possible in live view. Once focused, I exit live view and try to remember to change the ISO to the value that I’ll use for the photos. I wasted a half hour of my time last night because I forgot to change the ISO back. I would say that this is a lesson I won’t forget but I almost did it a second time last night when I changed to a different subject and refocused. (I’ll chalk it up to being sick–and me being outside taking photos last night for over an hour and a half speaks to how obsessed I am with this hobby.) I should also say that all automatic things on the camera have to be turned off or there is a risk of messing up the focus. For my very first set of exposures back in August, I left the image stabilization on thinking I would want as much help as possible with holding the camera still. But my stars were out of focus. I think it’s because image stabilization is designed to work in daylight photography.

Setting the exposure triangle correctly is the next challenging task. The camera has to be in manual mode so that you can correctly expose for the night sky. As a reminder, the exposure triangle is aperture, shutter speed, and ISO. I have researched my camera and found that the ideal ISO for astrophotography is either 1600 or 3200. This value gives the best dynamic range (range of colors that can be rendered in a single image without losing details) and signal to noise ratio in low light for my camera. I set the ISO to 1600. The widest aperture on my lens is f1.8 so I stop down a bit so that I have less vignetting in the corners of the image. I was also worried a bit about the light pollution where I was taking the pictures from. At home I would set the aperture to f2.4. Instead, I set it to f8. In astrophotography, the focal length of the lens determines the slowest shutter speed allowed before star trails start to form. I use the 500 rule to roughly determine the slowest shutter speed–500 divided by the lens focal length. For my 50mm lens, this rule tells me that the slowest I can go is 10 seconds but I know from experience that this will result in star trails. So I set the shutter speed to 8 seconds. Because I wanted to take 120 images, I used my intervalometer to automatically expose a number of images in a row without my intervention. I set the exposure time to 8 seconds, the interval to 10 seconds (which would leave 2 seconds between exposures), and the number of exposures to 30. Because the earth is rotating, the night sky moves which means my subject will move in the frame with each successive exposure. So I need to reframe the subject every so often. I have decided to do that every 30 exposures which means I will need to reframe 3 times in capturing my 120 frames.

Once the camera is focused and my exposure settings are set correctly, I point the camera at the subject. This can be challenging if the subject is not obvious in the night sky. I was shooting Pleiades and the California nebula last night so framing my subject was not hard. I chose to put Mars directly in the middle of the very bottom of the frame, a certain distance from the bottom. This put Pleiades on the right side of the frame and the California Nebula, I hoped, on the left side of the frame. I say “I hoped” because the California Nebula can’t be seen with the naked eye, even through a camera lens or telescope. But putting Mars in the center of the bottom of the frame also put Menkib, a bright star, in the center of the frame so I knew the California Nebula would be to the left of that.

Finally, I was ready to shoot. I started the intervalometer and it exposed 30 frames. I reframed the scene, trying to make sure Mars appeared in the same spot as when I first set the camera up. And I repeated this until I captured 120 frames. Because of what we have to do to get as strong a signal from our subject as possible, any noise in our exposures can cause significant problems. Noise comes from things like a dead pixel in the sensor, dust on the sensor, read out errors from the pixels, and so on. To minimize this noise, we take a series of calibration frames, each type of which deals with a different kind of noise. I won’t go into detail about what each kind of calibration frame does but after I captured my 120 images, I captured 25 dark frames. I should have captured 25 flat frames but I haven’t had a lot of success with taking them so I decided to skip them for the night. I have a library of bias frames which don’t need to be taken freshly every session.

I love being outside at night taking these images. But I think I love the next step, post-processing, even more. I’ll outline just the basic steps of what I have figured out works best for me at this point but I’m constantly learning about new techniques and software to use so this part of what I do is quite dynamic (which is what I love about it–always something new to try).

Here’s what a single exposure from last night looks like. Notice Mars in lower center of the image.

Not much to look at. If we brighten this image, we will make the subjects visible. But we will also brighten all that noise I talked about to the same level and the result will be quite horrible to look at. So we stack a set of images on top of each other to try to amplify the signal while minimizing the noise. That is, the goal of stacking is to maximize the signal to noise ratio. To stack, I used a free program called SiriL. I imported the raw image files from my camera into SiriL. I converted them to the .fits format, making sure to debayer the files in the conversion process. Debayering is the method by which the computer knows which pixels should be interpreted as red, which as green, and which as blue. I then apply my calibration frames to each of the 120 image files. (I skipped a step where I stacked each type of calibration frame to get a single stacked dark frame, a single stacked bias frame, and, if I had made them, a single stacked flat frame.) Once I had 120 calibrated frames, I registered them. Registering is a process in which the stars in each frame are aligned with each other (because remember the earth rotates so the stars appear to move with each exposure). Once all 120 frames were aligned with each other, I stacked them to create a single stacked image. SiriL saves this as a .fits file but GIMP, the software I use for my next step, sees .fits files as monochrome rather than color. So I also saved this single stacked image as a .tiff file which GIMP will see as a color image. Here’s what the .tiff file looks like (remember this is 120 images stacked on top of each other):

It doesn’t look much different than the single original image above. But now when I try to pull out the details of the subject, there will be far less noise. I next opened the .tiff file in GIMP, another free piece of software that is very similar to Photoshop. The first step in post-processing is to stretch the image. A non-technical definition of stretching an image is brightening the image so that the details of the image are visible. Here’s a first stretch of my image:

Notice the dark areas around the edges. Those are from the stacking process. Aligning the stars means that some areas around the edges won’t match and those register as these dark areas. So I cropped those parts from the final image. Notice also that the image is very green at the top and moves to a kind of greenish pink color as you scan down the frame. This is called a gradient and it made me wish I had done a background extraction in SiriL to try to remove the gradient. So I went back and did the background extraction on the registered images in SiriL and then restacked the images with background removed. After an initial stretch, the resulting image looked like this:

The gradient is gone which will make the rest of my post-processing easier. I won’t go into all the details of my post-processing in GIMP here. But I used levels, curves, and various filters to try to bring out the blue nebulosity between the stars in Pleiades and the red nebulosity of the California Nebula. Unfortunately, no matter what I did, I was unable to get the red from California. I think I just need more integration time. Integration time is the total amount of exposure time in the stacked image. For this session, I had 120 images of 8 seconds each which is 16 minutes. Probably not enough to bring out the colors of the nebula. I probably need at least an hour or two. I did get some of the blue nebulosity although I’d like to get more. Because I didn’t get the California Nebula, I decided to crop the image to reframe the objects of interest. Here’s the final image:

An interesting thing to notice about this final image is the spikes around Mars at the bottom of the image. These spikes are cause the diaphragm in the lens that control the opening and closing of the aperture. I think if I had used a wider aperture, these spikes would not be as prominent.

This is not a spectacular image in terms of photography but I’m proud of it because it demonstrates a huge amount of progress on my part. I still have a ton to learn about every single step in this process. I look forward to it!

The featured image shows my camera with a 14mm wide angle lens attached, taken at sunrise at my home.