Astrophotography Stacking Tutorial: Complete Processing Workflow from Raw to Final

Image Credit: NASA/ESA/Hubble – Deep field image showing galaxies billions of light-years away. Even Hubble uses stacking — it combined hundreds of exposures to create its famous deep fields. Public Domain.

Why Stack Images?

A single astrophoto exposure is noisy, faint, and underwhelming. But stack 50, 100, or 200 exposures of the same target, and something remarkable happens: signal builds up while noise averages out. The mathematical principle is simple — stacking N frames improves the signal-to-noise ratio by √N. Stack 100 frames and you get 10x better SNR than a single shot.

This is why astrophotographers spend hours collecting “light frames” on a single target. Professional observatories like Hubble and JWST do the same thing.

The Four Types of Calibration Frames

Before stacking, you need calibration frames to remove camera artifacts:

1. Light Frames

Your actual exposures of the target. These contain the astronomical signal plus noise, thermal signal, dust shadows, and optical vignetting.

2. Dark Frames

Exposures taken with the lens cap on, at the same exposure length, ISO, and temperature as your lights. These capture thermal noise (hot pixels, amp glow) that can be subtracted from your lights. Take 20-30 darks.

3. Flat Frames

Exposures of an evenly illuminated surface (a white T-shirt over the telescope opening lit by a dawn sky, or a dedicated flat panel). These map dust donuts and vignetting so the software can correct for them. Take 30-50 flats.

4. Bias Frames (or Dark Flats)

The shortest possible exposures with the lens cap on. These capture the camera’s read noise pattern. Take 50+ bias frames. Some workflows substitute “dark flats” instead.

Free Stacking Software

Deep Sky Stacker (DSS)

The most beginner-friendly option for Windows. Load your lights, darks, flats, and bias frames, click “Stack,” and it handles registration (alignment), calibration, and integration automatically. Output a 32-bit TIFF for further processing.

Siril

A powerful, free, open-source alternative that’s rapidly becoming the standard. Cross-platform (Windows, Mac, Linux). Offers scriptable workflows, gradient removal, photometric color calibration, and much more. Steeper learning curve than DSS but dramatically more capable.

ASTAP

Another free option with excellent plate-solving (automatic identification of the star field) and stacking. Great for batch processing.

The Processing Workflow

After stacking, the image will look dark and flat. Processing brings out the hidden detail:

1. Stretch the histogram: The most critical step. Your stacked image is linear — most data is compressed near black. “Stretching” redistributes the values to make faint detail visible. In Siril, use Asinh Transformation or Histogram Transformation. Go slowly with multiple gentle stretches rather than one aggressive one.
2. Remove gradients: Light pollution and optical imperfections create uneven backgrounds. Use Siril’s Background Extraction or the free tool GraXpert.
3. Color calibrate: Siril’s photometric color calibration uses star catalogs to set accurate, natural colors automatically.
4. Noise reduction: Apply luminance and chrominance noise reduction. Tools: Siril’s built-in denoising, or external tools like Topaz DeNoise AI or NoiseXTerminator.
5. Star reduction/removal (optional): Tools like Starnet++ (free, AI-powered) remove stars from the image, letting you process the nebulosity independently, then recombine.
6. Final adjustments: Curves, saturation, sharpening, and crop in GIMP, Photoshop, or Siril.

How Much Data Do You Need?

A general guideline for total integration time (total exposure across all stacked frames):

• Bright targets (Orion Nebula, Andromeda): 1-2 hours produces good results.
• Medium targets (most Messier objects): 3-6 hours for well-detailed images.
• Faint targets (galaxy details, faint nebulae): 10-20+ hours across multiple nights.
• Narrowband from cities: 10-30+ hours per channel for clean data through light pollution.

More data is always better. Many astrophotographers return to targets across multiple seasons, accumulating dozens of hours of integration.

Before and After

The transformation from a single raw frame to a fully processed stack is often dramatic. A single 2-minute exposure might show a faint smudge. Stack 100 of them, calibrate, and process — and you’ll have a deep, colorful image rivaling photographs in astronomy magazines. It’s one of the most rewarding experiences in the hobby.

Next in our series: Narrowband Astrophotography: Imaging in H-alpha, OIII, and SII