Siril Workflow: From Stacking to Final Image

Siril is a powerful free image processing software that has become a staple in the astrophotography community. Whether you are processing your first light frames or refining an advanced workflow, Siril provides professional grade tools without the price tag. This guide walks you through the complete process from raw frames to a finished image.

What You Need Before Starting

Before diving into processing, make sure you have these calibration and light frames ready:

• Light frames – Your actual exposure images of the target
• Dark frames – Same exposure time and temperature as your lights, with the lens cap on
• Flat frames – Short exposures of a uniform light source to correct vignetting
• Bias frames – Shortest possible exposures with the lens cap on

A minimum of 20 of each calibration frame is recommended, though more is always better. The calibration frames remove thermal noise, fixed pattern noise, and uneven field illumination from your light frames.

Step 1: Organizing and Converting Your Frames

Siril works with FITS files natively but can also handle RAW camera files. The first step is converting your files if needed and organizing them into the proper folder structure.

Create a working directory with subfolders for each frame type:

• lights/ – Your target exposures
• darks/ – Dark calibration frames
• flats/ – Flat field frames
• bias/ – Bias offset frames

In Siril, set your working directory using the toolbar or the command line. Siril will automatically detect frames in these folders when you begin processing.

Step 2: Preprocessing and Calibration

Calibration is where the magic begins. Siril applies your dark, flat, and bias frames to each light frame to remove systematic errors.

The calibration process works in this order:

1. Master bias is created by stacking all bias frames (median combine)
2. Master dark is created by stacking all dark frames, then subtracting the master bias
3. Master flat is created by stacking all flat frames, subtracting the master bias, and normalizing
4. Each light frame has the master dark subtracted, then is divided by the master flat

In Siril, go to the Calibration tab and select your frame types. Click “Start” and Siril handles the rest. You can monitor progress in the console log at the bottom of the window.

After calibration, inspect your calibrated frames. They should show reduced noise, even illumination across the field, and no hot pixels. If you see residual artifacts, you may need more calibration frames or your flat field source may be uneven.

Step 3: Registration (Alignment)

Registration aligns all your calibrated light frames so the stars match perfectly across every frame. Even with a tracking mount, frames drift slightly over time.

Siril offers several registration methods:

• Global alignment – Best for most deep sky images. Uses star matching to compute transforms.
• Three point alignment – Useful for wide field images with distortion from camera lenses.
• Comet registration – Aligns on a moving target instead of background stars.

For typical deep sky imaging, use global alignment. Siril will detect stars in a reference frame, then match and transform all other frames to align with it. The output is a set of registered (aligned) frames ready for stacking.

After registration, check the output. Frames with poor star detection or excessive drift are automatically excluded. If too many frames are rejected, your tracking may need improvement or your subframe exposures may be too long.

Step 4: Stacking (Integration)

Stacking combines all registered frames into a single image, dramatically improving signal to noise ratio. The more frames you stack, the cleaner your result.

Siril provides several stacking methods:

• Mean stacking – Simple average. Good for large datasets with minimal outliers.
• Median stacking – Rejects outliers. Best when you have satellite trails, hot pixels, or cosmic ray hits.
• Winsorized mean – A hybrid approach that clips extreme values before averaging. Often the best choice for deep sky.
• Rejection stacking – Uses sigma clipping to reject outliers statistically. Excellent for cleaning satellite trails.

For most deep sky work, Winsorized mean or sigma rejection stacking gives the best results. The stacking process may take several minutes depending on your frame count and image size.

After stacking, you will have a single linear FITS image. It will look dark and flat, with very little visible detail. This is normal. The real image data is there, but it needs stretching to become visible.

Step 5: Background Extraction

Before stretching, remove any gradients in your background. Light pollution, moonlight, and airglow create uneven backgrounds that become very visible after stretching.

Siril provides a powerful background extraction tool:

1. Open the Background Extraction tool from the Image Processing menu
2. Place sample points across the background areas of your image (avoid stars and the target)
3. Choose a polynomial degree (start with degree 2 or 3)
4. Click “Generate” to compute and subtract the background model
5. Review the result and adjust sample points if needed

The goal is a flat, even background across the entire field. If you see residual gradients after extraction, try increasing the polynomial degree or adding more sample points in the affected areas.

Step 6: Color Calibration

If you are working with a color camera (OSC), color calibration corrects the white balance so stars appear their natural colors instead of being tinted by your camera’s color response.

Siril offers two main calibration approaches:

• Photometric Color Calibration (PCC) – Uses catalog data to match star colors to their known values. This is the recommended method and works automatically for most images.
• Manual white balance – Adjust the red, green, and blue channels manually using the Histogram tool.

To use PCC, make sure your image has plate solving data (FITS header with RA/Dec). If your frames are not plate solved, Siril can attempt online plate solving. Click “Photometric Color Calibration” and Siril will match your stars to a catalog and compute the correct color coefficients.

After calibration, the background should be neutral gray and stars should show natural color variation (blue hot stars, yellow red giants, etc).

Step 7: Stretching (Histogram Transformation)

Stretching is the most impactful step. It transforms the linear data from your camera into the visually appealing image you see in astrophotography publications.

Siril provides the Asinh Transformation tool, which uses an arcsinh function to stretch the data without blowing out bright regions:

1. Open the Asinh Transformation tool
2. Start with a moderate stretch factor (around 10 to 20)
3. Adjust the black point to set the background level
4. Increase the stretch factor to reveal faint detail
5. Watch the histogram to ensure you are not clipping data on either end

For monochrome narrowband images, you will stretch each channel (Ha, OIII, SII) separately before combining them into a color image using the Channel Combination tool.

Take your time with stretching. Small adjustments make a big difference. The goal is to reveal faint nebula detail while keeping star cores from bloating and maintaining color in bright regions.

Step 8: Final Processing

After stretching, apply final adjustments to polish the image:

• Deconvolution – Sharpens stars and fine detail. Use the Deconvolution tool with a small regularization value to start. Overdoing deconvolution creates ringing artifacts around stars.
• Noise reduction – Apply gentle noise reduction to smooth the background without losing detail in the target.
• Curves and levels – Fine tune contrast and brightness using the Curves tool.
• Color saturation – Boost saturation slightly to bring out nebula colors.

Each of these adjustments should be subtle. The best astrophotography processing is invisible. You want the viewer to see the object, not the processing.

Step 9: Export Your Final Image

Once you are happy with your processed image, export it for sharing:

• TIFF (16 bit) – Best for further editing or printing. Preserves the full dynamic range.
• PNG – Good for web sharing with lossless quality.
• JPEG – Smallest file size, acceptable for social media and web posts.

Use File, Export Image in Siril and choose your format. For web publishing, export as JPEG or PNG at full resolution, then resize to your needs.

Tips for Better Results

• More frames always beats more processing – A clean stack of 50 frames will always look better than a heavily processed stack of 10 frames.
• Calibration frames matter – Do not skip darks and flats. They fix problems that no amount of processing can recover from.
• Save intermediate results – Save your stacked image before stretching so you can try different processing approaches later.
• Use Siril scripts – Siril supports scripting for repetitive tasks. Once you find a workflow that works, automate it.
• Process non destructively – Keep your linear (pre stretch) data. You can always start over from the stack.

Conclusion

The path from raw frames to a finished astrophotography image follows a clear pipeline: calibrate, register, stack, extract the background, calibrate colors, stretch, and finalize. Siril handles every step of this pipeline with professional grade algorithms, and it keeps getting better with each update.

The most important thing is to get out and collect data. No amount of processing can replace good data. But with Siril in your toolkit, you have everything you need to turn that data into something beautiful.

Clear skies!