PROCESSING BASICS

How to stack astrophotography frames

Why one exposure looks like noise, why forty of the same exposure don't — and the four-step workflow that turns raw frames into a real photo.

By the Astro Easy Mode pipeline team · updated 2026-08-02

If you've pointed a camera or telescope at the sky and opened the resulting file, you've probably had the same experience: it looks like static. A faint smear, maybe, if you squint — but nothing like the glowing nebula photos you were chasing. That gap between "raw frame" and "finished photo" is almost entirely closed by one process: stacking.

What stacking actually does

Stacking means combining many exposures of the same target into one image, so that the real signal (starlight, nebula glow) reinforces itself while random noise (sensor grain, thermal noise, sky glow variation) averages itself out. A single 60-second exposure of a faint nebula is mostly noise with a whisper of signal buried in it. Fifty of those same exposures, stacked, keep the signal — which is consistent frame to frame — while the noise, which is random each time, gets diluted. The math is straightforward: noise falls roughly with the square root of the number of frames, so going from 1 frame to 10 frames helps far more than going from 40 to 50. That's why even a modest stack of 20-30 frames from one clear night already looks dramatically cleaner than any single sub.

It's worth being precise about what stacking is not: it doesn't invent detail that isn't there, and it doesn't brighten the image on its own. It only improves the signal-to-noise ratio. The image still needs a final step — the stretch — before it looks like a photo at all. More on that below.

The standard workflow: calibrate → register → stack → stretch

Every stacking tool — free or paid — runs roughly the same four stages under the hood.

1. Calibrate

Calibration subtracts sensor artifacts using extra reference frames: dark frames (exposures taken with the lens capped, at the same settings, capturing the sensor's own thermal noise and hot pixels), flat frames (exposures of an evenly lit surface, capturing dust shadows and vignetting so they can be divided out), and bias frames (the shortest possible exposures, capturing the sensor's baseline readout noise). Skipping calibration frames isn't fatal — you can still stack light frames alone — but the final image will carry more of these artifacts into the stretch.

2. Register (align)

Even with a tracking mount, stars drift a few pixels between exposures — tracking error, polar alignment drift, atmospheric refraction. Registration detects stars in every frame and rotates/shifts each one so the stars line up precisely on top of each other across the whole set. Without this step, stacking would just blur everything into a smear.

3. Stack (integrate)

With frames calibrated and aligned, the software combines them pixel by pixel — usually averaging or a smarter statistical method (like sigma-clipping, which throws out outlier pixels such as satellite trails or cosmic ray hits before averaging the rest). The output is one "integrated" image with a much better signal-to-noise ratio than any single input frame, but it will still look almost entirely black.

4. Stretch

This is the step that turns a black frame into a photo. Astronomical data has an enormous dynamic range — the faint outer glow of a nebula and a bright star core can differ by a factor of thousands — so the raw stacked data is compressed into a narrow, dim sliver of what your screen can display. Stretching remaps that data (typically with a histogram or curves adjustment) so the faint signal becomes visible without blowing out the bright cores. It's the single step beginners underestimate most: a technically perfect stack still looks like nothing until it's stretched correctly.

The three tools people actually use

None of these are bad tools — each is well-regarded within its niche, and plenty of excellent images come out of all three. The honest differentiator for a beginner is time and tolerance for troubleshooting, not raw capability.

When to DIY, and when not to

Learn to stack yourself if: you enjoy the craft side of the hobby as much as the photography, you want full manual control over every parameter, or you plan to do this repeatedly across many sessions — the time invested in learning Siril or PixInsight pays off across dozens of future images.

It makes more sense to have someone else process it if: you have one good session's worth of frames sitting unprocessed and just want to see the result, you're time-poor and the software wall is genuinely keeping you from ever opening the files, or you've tried DeepSkyStacker or Siril and hit a wall that killed your enthusiasm rather than building it. There's no shame in this — plenty of people who are excellent at capturing data have zero interest in spending a weekend debugging a stacking pipeline, and that's a completely reasonable trade to make.

Frequently asked questions

More guides

• DeepSkyStacker vs Siril: which one for beginners?
• Why do my astro photos look like noise?
• Smart telescope vs DSLR for astrophotography
• First astrophotography session checklist
• All guides
• Astro Easy Mode home