Stacking: How Amateurs Beat Light Pollution with Math

Modern amateur astrophotography leans on one statistical trick: stack many short exposures instead of taking one long one. The signal adds; the noise partially cancels. That single idea is why $2,000 rigs from suburban backyards routinely produce images that would have required an observatory-class telescope thirty years ago.

Signal, noise, and why averaging helps

Every exposure captures three things: signal from your target, signal from light pollution and skyglow, and random noise from the camera. When you average N frames pixel by pixel, the target signal stays put, the light-pollution offset stays put (until subtracted separately), and the random noise drops by a factor of √N.

That's the whole trick. Stack 100 frames and noise is reduced 10× compared to any single frame. Stack 400 frames and it's 20×. Diminishing returns eventually, but you can push signal-to-noise ratio arbitrarily far with time.

Why not just take one very long exposure?

A single 10-minute exposure has three problems short exposures don't. First, satellite trails, airplanes, or a passing cloud ruin the entire frame. Second, tracking errors are cumulative — one bump and the whole exposure is smeared. Third, in bright suburban skies, the sky itself saturates the sensor well before you've collected enough signal from a faint target.

Stacking 60 frames of 10 seconds each gives you the same total exposure time (10 minutes), but a bad frame just gets tossed, tracking errors don't accumulate across frames, and the sky background never saturates. It's fault-tolerant in exactly the ways one long exposure isn't.

Calibration frames: the boring but critical part

Real stacking uses four kinds of frames:

Each calibration type is averaged into a 'master' frame that gets subtracted or divided out during processing. Skipping calibration is a common shortcut, and it costs you real depth — donut-shaped dust shadows and warm-pixel constellations will fight you at every stretch.

A realistic workflow

1. One night, start to finish
2. 01Cool the camera (for dedicated astro-cameras) to −10 °C. Take 20 darks at that temperature and your target exposure time.
3. 02Take 100+ light frames of your target through the night. Guide if possible. Discard any obviously bad ones as you go.
4. 03At end of session, take 30 flat frames — point at an evenly lit surface (twilight sky, illuminated white t-shirt over the objective) and adjust exposure so the sensor is mid-tone. Same optical path as your lights.
5. 04Load everything into stacking software: DeepSkyStacker (free, Windows), Siril (free, all platforms), or PixInsight (paid). It auto-aligns, integrates, and applies calibration.
6. 05The output is a linear, low-contrast master frame. Bring it into an editor (Siril, PixInsight, GIMP with astro plugins) and stretch, remove gradients, and adjust color.

Why stacking beats light pollution

Under Miami suburban skies (Bortle 7–8), a single 60-second exposure is dominated by orange sodium/LED skyglow. But that skyglow is nearly constant — it can be measured and subtracted. What varies frame to frame is random noise, and stacking averages that away. Add a narrowband filter (like Optolong L-Extreme) that passes only H-alpha and OIII light and blocks everything else, and even inner-city imaging of emission nebulae becomes possible.