Photographing Star Trails – How to shoot and stitch stars.

Two things to throw out there first: 1) I am an average amateur photographer at best (so don’t trust me… even though I’m an engineer). 2) It’s been far too long since my last post (due in no small part to school work). Here’s what I’ve been working on.

For the months now (ever since the Perseids fiasco), I’ve been meaning to catch a proper meteor shower. The last time I tried, I was facing the wrong not-entirely-optimal direction and staring into Detroit’s light pollution. This time around, I went all-out. The conditions were supposed to be favorable, with no moon to blind you. I drove down to a park with a long, east-facing field so I could see more of the horizon, dressed in super warm clothes, and was ready to camp there overnight in the car.

And that’s when I realized the “20-30% cloud cover” that predicted really meant “100% thin cloud coverage.” I setup the timelapse anyways, making good use of my tripod’s independently deployable feet and setting it up to wedge my camera against the top edge of the passenger side door for stability. I rolled down the window just a hair to eliminate the potential for fogging as I slept and popped in noise isolating earbuds to block out the sound of the shutter.

I woke up 2 hours later freezing. Apparently PJs under jeans under sweatpants was too thin for below-freezing temperatures, as were my thick gloves. The exposed bit of my nose was frustratingly cold, and compelled me to grab the scarf off my camera. I had been using it to help keep the battery compartment from freezing (I had also intended on stashing a few hand warmers in there, never bought them). Woefully under-dressed and miserable, still seeing nothing but a formless grey sky, I packed everything up and drove home. This was at around 3:30 AM. I was convinced that I had captured nothing of value and considered formatting my card at once, but I was too tired to follow through.

After waking up, eating breakfast, and paying the Makerbot Thing-o-matic in the library a visit (project to be explained later), I got back to my apartment and finished cutting together the montage seen in the beginning of this post. While I was waiting for it to render, I decided to process the RAWs I had taken the night before. Just to see what I could come up with. Maybe with some flagrant abuse of pseudo-HDR the clouds might look kind of cool. A few images later, I saw a streak on one of the processed images. It had a strange diamond profile, and none of the dashed strobe lines that a normal plane had. And I was really glad I hadn’t just erased everything on the card. I canceled my render and decided to toss in my Geminid footage as another instance of almost-fail.


To this day, I still don’t know what the best camera settings are to capture meteors. If you use a long exposure (30 seconds or more), then you risk having a really dim meteor streak since the stars have had so long to expose. Do it the other way around, and you face the problem of potentially missing it with such a short exposure time, unless you compensate by taking a bajillion pictures with short intervals in between. I opted for something in between, 20 second f/3.5, ISO 800, interval 40 seconds. That gave me 50% coverage and enough time that I wasn’t worried about overflowing buffers or burning through my battery.

The program I used to achieve the star-trail effect is called StarStaX, a bit of awesome freeware by an awesome German dude (official title). Basically you load in pictures, select a blending mode, optionally add in a noise profile (pic taken with a lens cap on), and out pops magic. The result is a star trail image which you can save. To get a video, you tick the check box to save an image at every step of the overlay process, and then make a video sequence with those.

To de-flicker the “Camera Endurance Testing” footage, I found a free plug-in (MSU Deflicker) for the GNU program VirtualDub and ran finalized footage through it. It’s not nearly as smooth as a commercial solution like LRTimelapse or GBDeflicker… but it’s free. And you can see a clear difference compared to my “Perseids” attempt.

The Perseids-lapse I took with the camera on full manual, but at f/5 or so. I thought that shooting less-than-wide-open might sharpen things up since it was hard to focus to infinity without over-focusing. Instead, there were minor variations in the set aperture each time my camera took a picture, resulting in flicker. The “Camera Endurance Testing” footage was taken on full auto, which is why things weren’t instantly washed out when the sun rose. And FYI, I was testing the endurance of the Chinese knockoff extended battery grip for the D5100. Got about 5-6 hours of long exposures at reasonably chilly temperatures.

Alright. I know I’ve been rambling about random bits of information for far too long now. In short: StarStaX I will definitely be using again, shoot at max aperture (your camera will thank you), and dress SUPER ULTRA MEGA warmly if you’re going to be outside for more than an hour.

Leave a Reply