Stephen Bay's Photography Blog

Image Stacking vs Tracking for Landscape Astrophotography

In June, I was out in Anza Borrego and managed to get my first real test out of using a star tracker. Previously, I had been stacking images with Starry Landscape Stacker (SLS) to reduce noise. Given that I had some extra time, I took some shots to use in comparison and thought I’d share them here with my observations.

Test Setup

  • Sony A7R2, 16-35 FE f/4 lens
  • camera -> BH-55 ballhead -> iOptron Skytracker Pro -> equatorial wedge -> gitzo 3541 LS
  • 6 stacked shots @ ISO 800, 16mm, 20s exposure, f/4 taken with tracker off
  • 1 tracked shot @ ISO 800, 16mm, 120s exposure, f/4
  • temperature was in mid 70s
  • images recorded in Sony uncompressed RAW format

Stacked images were averaged in Starry Landscape Stacker and preprocessed according to SLS guidelines (contrast -100, noise reduction off, sharpening off, CA on, no profile corrections, exp +3.6). Default sharpening and color noise reduction was then applied to the tracked composite.

The tracked image was processed as follows (contrast -100, exp +1, CA on, no profile). This resulted in two very similar images in terms of overall exposure, histograms, etc.

Note #1: The sony suffers from the star eater effect and this effects both stacked and tracked (there is a somewhat convoluted way to turn this off, but I didn’t do so for the test).

Note #2: I would have used a smaller ballhead than the RRS BH-55 on the tracker but I forgot it.

The test setup

Comparison Images

Stacked is always on the left, tracked right. Click on the images to view them full size.

The first image here is the comparison of the whole frame: The test setup

Center of the image with the core: The test setup

Planet with sunstar: The test setup

Left edge of the frame, middle vertically: The test setup

Top left corner: The test setup

Keep in mind these are 100% crops of a 42mp camera. On a HiDPI / retina screen (220ppi or so), this corresponds to looking at a 24”x36” print. On a regular screen, it would be something like 48”x72”.

Observations

  • There are some slight differences in WB. The stacked shots are greener whereas the tracked shot is more magenta.
  • The tracked shot is sharper everywhere. Very noticeable especially in the dust clouds in the MW core and in the sunstar for the planet.
  • Noise levels appear to be visually about the same. This is as you might expected given that the total exposure time is the same.
  • Tracked shot has more pin-point stars. The rule of 500 suggests an exposure of 31s for stacked shots. I used a more conservative 20s but the stars still look like hyphens.
  • Stacking with SLS automatically eliminates hot pixels in the sky. The tracked shot has a quite a few and I didn’t think to take any dark frames (because I have always been stacking up to this point).

My Conclusion

The stacked and tracked shots have similar noise levels but the tracked shot is significantly sharper and has more pleasing stars (no trailing). If I can ever get good at doing alignment, I’ll stick to the tracker. But when I’m pressed for time, or it’s freezing cold, I’ll use stacking. Note for tracking you have to blend a sky shot with a separate foreground. Programs like SLS do the blending for you.