↓Skip to main content

Siril 1.4.4 DSO editing process

Table of Contents
It’s a good idea to Save As after every edit, so that it’s easy to return to a previous step if something goes wrong.

Exposure Settings #

Light exposures #

  • ISO 1600-3200
  • Check the exposure that histogram peak is at most about 2/3 way to the right
    • Most of the detail is captured at mid-brightness
  • Don’t use ISO under 1600: (STN) Signal to noise degrades fast.
  • Increasing ISO above 1600 will not give any advantage. Optimal signal-to-noise ratio is at about ISO 1600-3200.

Calibration Shots #

Bias Dark Flat
Description Images, with the lens cap on Images, with the lens cap on Gray-white pictures, but not blown out
# of images 50-100 20-50 20-50
ISO Same* Same* -
Exposure time ~1/4000 s Same* -
Temperature - Same* -
Optics Linssinsuojus Linssinsuojus Same*

“Same*” = Same as for light exposures

Re-using Calibration Frames #

You can re-use calibration frames, as long as conditions have not changed much. I use different calibration frames for different exposure lengths and temperatures.

Light frames may change with dust in optics or sensor.

1. Stack #

Scripts #

File structure #

The scripts assume the following file structure:

+ folder
	+ biases
		NNNN.CR2
		... 
	+ darks
		NNNN.CR2
		...
	+ flats
		NNNN.CR2
		...
	+ lights
		NNNN.CR2
		...

1a Siril Script #

Script OSC_Preprocessing.ssf will do all steps at once.

1b With my scripts #

You can download my scripts that I use to process DSLR (color) images.

The Scripts split the process into the following steps.

  1. Convert calibration images to .ser
  2. preprocess calibration images
  3. convert light images to .ser
  4. calibrate and align lights
  5. stack final image as .fit

Each step may take several minutes, depending on your hardware.

Combination scripts 0123 and 01234 do the corresponding steps in one go.

2. Crop Stack Artifacts #

The edges may have strong noise and color artifacts which will scew later gradient and noise removal steps.

  1. Draw a rectangle around the are you want to keep
  2. Right click Crop > Crop

3. Background extraction #

3a Siril Native #

Use autostretch to see what you’re doing.

Background extraction is in Image Processing > Background Extraction

  1. Settings
    1. Method: RBF (default)
    2. Smoothing: 0.00-0.5
      1. If the backround doesn’t become flat, reduce smoothing until it does.
      2. Add samples per line if necessary
    3. Samples per line: 20-40 (something reasonable)
    4. Grid tolerance: 2
    5. Correction: Subtraction
  2. Click Generate
  3. Remove sample point that are on the target or bright stars - i.e. not backround - by right clicking
  4. Compute Backround
    1. Change settings if necessary and comput again
  5. Apply
Remember that while auto streatch is on, any residual bacground pattern is amplified and the picture may appear worse, since the auto streatch will over streatch what’s left.

3b GraXpert #

Background gradient #

Use the GraXpertAI.py script available from Get Scripts menu.

Use the background removal first. You may consider reducing the smoothing for complex pattern gradients.

I have good experience with smoothing as low as 0.0.

Noise Reduction #

It’s a good idea to use the noice reduction at this point. The GraXpert NR is (as far as I know) designed to be used on unstreatched data. It’s also pretty good.

4. Color Calibration #

4a Photometric Color Calibration #

Platesolve #

Tools > Astrometry > Image Plate Solver

  1. Search the name catalog number of the target (e.g. “M27”)
  2. Prefer Simbad over NED

Calibration #

Image Processing > Color Calibration > Photometric Color Calibration

  1. Crop out different colored fringes so they won’t throw off the calibration
  2. Optionally expand Backround Reference
    1. Draw a box on empy piece of sky, uncheck Automatic detection
    2. Click Use current selection
  3. Click OK

4b Local Catalogs #

If online catalogues don’t work, you can install catalogs locally and use them. Python Scripts > Core > Catalog Installer.py

After install, try [step 3a](#3a Photometric Color Calibration) again.

4c Using basic Color Calibration Tool #

  1. Draw a box around an empty piece of background
  2. Click use…
  3. Draw a box around the target
  4. Click use…

4d Using histograms and streatching #

If you use streatchig to adjust colors, before star removal, you’ll need to do an initial streatch too, before star removal. You’ll also need to uncheck “linear data” in star removal step.
  1. Use Generalized Hyperbolic Streatch Transform
  2. Uncheck Logaritmic scale
  3. Use zoom on the histogram 1-60x
  4. Select one color at a time (colored circles) and try to align the peaks
  5. If a peak is much wider than other peaks, you can widen other peaks like this
    1. Add streatch factor to widen the peak
    2. Use symmetry point to move the peak back aligned with others
    3. Use Highlight protection to protect the highlites

5. Star Removal #

Setup #

  1. Download Starnet CLI
  2. Open Open Scripts > Get Scripts
  3. Search “starnet”
  4. Put a checkmark next to StarNet.py
  5. Apply

Use The Script #

  1. Scripts > Python Scripts > Star Net Astro > StarNet.py
    • (if starnet2 wasn’t found, set the path with SetPath button)
  2. If you haven’t streatched the data, check Processing Options > Linear Data
  3. Check Output Options > Unscreen stars
  4. Run Starnet
In my testing with Siril 1.4.4 and Starnet 2.5.4, I got the best results with Unscreen stars, but Siril star layer (substraction) may also be worth a try. You can select multiple outputs and see which works for you.

6. Selective Saturation (optional) #

Image Processing > Color Saturation

  1. Add saturation to your liking. (0.5~1.0)
  2. Zoom into an edge area and use the Min and Max sliders at the bottom of the screen, so that you can see the faintest nebulosity
  3. Adjust background factor so that the faint parts of the target get a boost, but the backround color noise doesn’t increase

7. Stretch #

7a Generalized Hyperbolic Streatch Transform #

Open Image Image Processing > Streatches > Generalized Hyperbolic Streatch Transform

Initial streatch #

Make sure you are in Linear mode and the Min/Max sliders are at their limits. Typically 0 and 65535 on a 32 bit image
  1. Slowly start adding streatch factor, until you can see your target cleary
  2. Use Highlight protection to stop highlites from growing uncontrollably and to preserve highlight details and contrast.
  3. Apply

Second streatch #

Use Logarithmic scale in histogram to see the location of the peaks more easily
  1. Draw a box around the faintest feature you want to preserve
    • Use the sliders at the bottom of the screen to help you see the faint detail, but remember to normal after you’ve found the detail.
  2. Click the dropper at Symmetry Point slider
  3. Add Streatch Factor moderately, such that the dim feature is sufficiently visible
  4. Add highlite protection to prevent the target from getting too bright in the process
    • Adjust such that the histogram stays approximately centered
  5. Adjust Local streatch intensity to increase contrast at the faint features edge, bring the faint features from the backround.
  6. Nudge symmetry point left until no more faint features come out
  7. Repeat this step untill all the waint areas are sufficiently visible

Final streatch #

  • Adjust the final brightness as you like
  • You can use the symmetry point to focus adding contrast to parts that are lacking it.
  • You can use the highlite and black point protection to help keep preserve detais

Adjust black point #

|650

  1. Select Linear Streatch (BP shift) from the Generalized Hyperbolic Streatch Transform tool. (Type of streatch dropdown at the top of the window)
  2. Adjust the background sky darkness. to your liking
Do not make the sky completely black at this point. A crushed black becomes a problem if you want to do any more fine edits to the picture, such as noise reduction.

Adjust Saturation (Streatch) #

Adjust the symmetry point to find the point where the backgound noise isn’t boosted.

This behaves similarly to the saturation tool.

7b VaraLux Hypermetric Streatch #

Using the VaraLux Hypermetric Streatch script, you can get a good stretch in one step.

The script is available from Get Scripts menu.

Select: Scripts > Python Scripts > VeraLux > VaraLux_HyperMetric_Streatch.py

  1. Select your camera sensor or leave on default
  2. Click Live Preview
  3. Adjust Target Bg to set the desired background darkness
  4. Click Auto-Calc Log D (optional)
  5. Click Process

8. Recombining #

Select: Star Processing > Star Recomposition

Adjust Starless and Star Mask layers

During recombining black point adjustments can be made to each layer, as well as light stretch, if necessary.

9. Final Touches #

|350

Filters menu has numerous useful tools for polishing the image, such as

  • Deconvolution
  • Noise reduction

and various other filters to address specific issues.

Cropping #

You may crop the image, by drawing a rectangle around the desired area, then right clicking the image > Crop > Crop

10. Exporting #

To Save the file with a new name, click the disc icon shown above.

Depending on what you want to do with the file, you may choose from a number of formats. Unless you have a particular software that needs a different format, I recommend PNG as a good middleground.

Editing in other software #

You may want to edit the image further in other software.

Gimp #

Gimp is similar to Adobe Photoshop in that it has very powerful tools for finer edits.

GraXpert #

GraXpert is a powerful tool to remove gradients and noise in .fit images.