I wrote a while back about software to generate star trails from a series of images of the sky. They are pretty and a fun and interesting form of art, but I wanted to have a go at taking photos of the night sky as it actually appears: still, and without trails.
It’s pretty hard to achieve good photos with a consumer DSLR. You have to keep the shutter speed faster than 4-5 seconds, otherwise you start to see motion blur creeping in. This means you have to turn up the ISO sensitivity of your sensor, which in turn leads to noise. As soon as you start to examine these photos closely, the noise is so distracting and it ruins the picture.
There are some ways to alleviate this – buying a faster lens, or a lower noise, higher sensitivity sensor – but these will cost you money. Lots of money.
I’ve been playing with the concept of stacking – where you take multiple images and later combine them to remove the noise. “Adding” the photos is the easy bit, but the stars will have moved slightly in between each photo, so they will need to be aligned by rotation and translation.
Some crazy people have attempted to do this by hand in Photoshop, but I’m far too impatient for that. There are some apps for Windows (both commercial and free) that will automatically align and stack photos. However, I wasn’t able to get any of them to work reliably under Wine on Linux, so I set about finding a native Linux stacking app.
The best I found was ale, which is command-line only but very easy to use. Both Fedora and Ubuntu package ale by default. To use it with default settings, it’s just a case of doing this:
ale image1.jpg image2.jpg image3.jpg image4.jpg output.jpg
This picture of The Plough was taken using a Canon 450D with 50mm f/1.8 lens. I took 12 exposures at ISO1600, 1″, f/5.0 and stacked them with ale. On my PC, a Core 2 Quad 8200 with 8GB RAM, it took about 53 minutes to stack 12 photos.
The stars are very small so you will probably need to click to see the larger version. You’ll notice that the background is very dark as ale has removed sensor noise, cosmic ray strikes and light pollution (I’m shooting in a city).