After my last blog post on my recent trip to Idaho, a few folks were curious about how I created the star trails image in the post. So I thought I’d share a bit on how it was done. First, here is a movie of the actual image (I’ll explain what that means later):
The image is basically an image of star trails through the night. Digital photographers have been creating star trails images since long before I picked up a camera, so I was able to learn from some of the literally thousands of examples on the internet. I’d even done some star trail shooting already. In fact, one of my most popular images is a star trails shot I did a few years ago in Lake Tahoe. This time around, I wanted an image with some of the landscape (trees, etc) exposed and trails showing the path of the stars in the night sky. For the landscape, that meant I would need to expose enough to at least get some of the tree detail, and the only available light would be moonlight. For the stars, that meant I would have to shoot for a very very long time to track them through the sky. I also knew that I didn’t want the moon in the shot.
What I Needed
- a camera with fully charged batteries….its a lot of shooting no matter how you do it
- a tripod…no human can stay perfectly still for very long
- a cable release to act as my shutter finger, holding the shutter button down for hours
- very minimal light pollution…this is where being out in the middle of nowhere (and at higher altitude) really helps
- knowing the position of the North Star….this is optional and allows you to position the camera to capture star trails circling Polaris
- knowing how the moon would travel through the sky overnight (less optional)
- and most out of your control, a clear sky
Research-lite
The night before I did the shoot, I took a look at where the moon was in the sky and how it traveled. Fortunately, the moon was in the southern sky. So, I’d be able to shoot toward the North Star without the moon ever coming close to being in the picture, AND I’d be able to use the moon as a light source to illuminate the trees in the foreground. Here is a really rough diagram of what this looked like:
Shooting
I’ve shot star trails in two ways. Both of them take a long time. I’ve never really done any math to figure out how much stars will move through the sky in a given time. Usually I just shoot for a while (30-90 minutes) and see what I get. There are two ways I know of to shoot them: 1) take a single image or 2) take lots and lots of images and then combine them (I use photoshop to do that part). There are advantages and disadvantages to each, and it makes more sense to talk through those as I describe the techniques (mostly because thats how I learned about them).
I decided I would try the single shot approach on the first night (and leave the North Star out of the image). With a single shot, you basically leave the shutter open for a really long time (this is where the cable release comes in) and capture the stars moving through the sky. I decided to try 60 minutes. But at what settings? I shot using a 16mm lens, and because I wanted to focus on the stars, I set my camera to manually focus and then set the focus to infinity. I like to shoot at a mid-range aperture (5.6-8), though there’s no real scientific process for choosing. I always do a test shot at 30 -60 seconds just to make sure that I am picking up the stars in the shot. If not, I adjust the ISO to increase the sensor’s sensitivity and/or open the aperture a bit. I finally settled on the following settings to do my first hourlong shot: f/8.0, ISO 800, 16mm
I used my cable release to open the shutter. That was at 12:30am. I went into the house and set an alarm for 1:30am, because I’d need to come back out and close the shutter. Downside #1! Without a device called an intervalometer, I needed to come back to the camera to close the shutter and stop the sensor from capturing light for that frame. HUGE pain, especially in the middle of the night. At 1:30am, I trudged back outside (to coyotes howling, no less), closed the shutter, and saw this on the LCD:
Fail! It looked like high noon! That moon was SO much brighter than I bargained for. It’s hard to believe that this much light could come from the moon….but then again, the shutter was open for an hour. At this point, I needed to either go get some rest and fight another day or tweak my settings and commit to staying up for at least another hour. You can probably guess what happened. So…after doing a quick test using f/11 & ISO 160, I opened the shutter again. At 2:30am, I came back down to see a much better image:
I liked this image much better at first glance, but that leads me Downside #2 of shooting star trails digitally at very long shutter speeds: noise. And lots of it. Here is what the image looks like zoomed in:
You can see lots of red, green, and blue dots. Those are pixels. The issue is that when shooting at long shutter speeds, heat from the sensor causes “thermal noise” which manifests as the colored pixels you see above. You can go crazy learning about this by Googling “long exposure noise”. There are ways to deal with that noise in post processing, but I am not a fan of those because its really difficult to effectively reduce noise without reducing detail. I also didn’t like the fact that the trees were complete silhouettes. I decided to give it one last try. I opened the shutter one last time and went to bed, assuming that my camera battery, which was at about ⅓ charged at that point, would die before sunrise (and close the shutter with it). It didn’t. When I came back down 5 hours later the shutter was still open, the sun was blazing right into the lens and the resulting image was a big bright pure white rectangle. Oops.
Night #2
On the second night, I decided to try my other approach: shooting lots of consecutive images through the night and combining them to draw out the star trails. I got set up well before sunset. I already knew where the moon would be, what time it would set, and what a good composition looked like. I set up the camera to shoot 30 second exposures at f/6.3 and ISO 4.0 and took a test shoot. Looked good. After I opened the shutter at 12:39am (by putting the cable release in the “locked on” position), I just went to bed. By the time I woke up at 5:15, the camera had taken over 500 30-second exposures and the sun was rising. I unlocked the cable release to stop the camera from taking more pictures. By the way, the camera, the lens, the tripod…everything….was dripping with morning dew. Small price to pay, I say. Here is one of the images, taken at 12:53am:
Believe it or not, even a shutter speed as short as 30 seconds reveals that the stars are actually moving relative to the earth. Zooming into the upper right corner of the image, you can tell that stars are moving:
All that was left was to combine the images. The great thing about this approach is that the image doesn’t really get “cooked” until I combine the individual images in photoshop. So I could use the first 100, the last 200, or whatever consecutive set of images I like to make the final version. I used the first 450, since the last 100 or so were too bright with the sun already rising. The movie above contains the 450 images I used and gives a sense for how smoothly you can capture the movement of the stars and the moon’s light. Here it is again, at 24 frames per second:
My basic approach to combining the 450 images in photoshop could not be much easier: open each image as its own layer within one file, select all layers, then set the blend mode to Lighten. Here is one site that explains the technique in more detail, but it really is about that simple. After doing that, I tweak the basic items like color temperature, saturation, and contrast to arrive at something I like. Even though that takes lots of time, this approach to shooting star trails gives me TONS more flexibility in the final image than a single exposure, because you’re free to create as many version as you like using different sets of consecutive images. Here is the final image:
If you’ve made it this far, thanks for indulging me.