Peter Oshkai photography

My current workflow for processing one-shot colour images of deep-space objects, particularly nebulae, heavily relies on the process of reconstruction a foraxx palette outlined in this video by Paulyman Astro (https://youtu.be/Gy42AeZ_XB4?si=lNb4B7a0jVKWi4I1).

The video is very detailed, and it has been exceptionally useful for me, but I found myself scrolling through it and pausing so much that I wanted to have a written summary of the steps that I could quickly refer to. Apologies in advance, if the following steps appear out-of-context – they are really short notes that would make sense to those who have used the Pixinsight software and worked with the foraxx palette.

Before this image processing is even started, I generate an integrated image using the following pre-processing sequence (see this guide by Adam Block for details: https://youtu.be/VKOTCuqD2Qs?si=EdbONwT8GO_DUAkR) :

First of all, all the individual exposures (light frames), which are typically 5-min long each, are calibrated with “flats”, “dark flats” and “darks” using the WeightedBatchPreprocessing script. 

Then, the resulted de-bayered images are aligned using the StarAlignment process.

Finally, the aligned images are integrated using the ImageIntergation process with Winthorised Sigma Clipping background rejection method. This produces the “integration.xisf” image, which is the basis for the nebula processing workflow itself:

1. Right-click on the identifier tab at the top-left of the image frame and set the new identifier to ‘osc’ (for ‘One-Shot Colour’).
2. Use Process>All processes>ScreenTransferFunction to preview a stretched image. I unlock the RGB channels before pressing the “nuke” button to avoid a high colour cast. This is not important though, because nothing is actually being done to the image – this is just a preview.
3. Use Process>All processes>DynamicCrop to crop the image.
4. Use AutomaticBackgroundExrtactor with Function degree (under Interpolation and Output) set to 1 and Target Image Correction to ‘Subtraction’. This works if there is simply some light pollution gradient in the image and little to no vignetting. Otherwise, use the DynamicBackgroundExtraction process.
5. Split the RGB channels using an icon in the toolbar.
6. Set the identifier of the Red channel to ‘ha’ for “Hydrogen alpha”.
7. Combine the Green and the Blue channel using the following Pixelmath expression: 0.5*B+0.5*G. Give the resulting image an identifier ‘oiii’. 
8. Use StarExterminator process to extract the stars. Check “Generate Star Image” and “Unscreen Stars” boxes. Drag the triangle from the process window onto the ha window and then onto the oiii window.
9. Use Generalized Hyperbolic Stretch (GHS) script or process to stretch the ‘ha’ and the ‘oiii’ (starless) images.
   • First stretch: Just right of the peak. Local stretch intensity ~10. Stretch factor ~3.
   • Second stretch: Secondary drop-off (log view). Local stretch intensity ~5.
10. Foraxx process: Create the false Green channel (‘ho’) using the Pixelmath expression: (ha*oiii)^~(ha*oiii). Press the square button.
11. Boost the brightness of the ‘ho’ image by doing the first-level stretch in GHS.
12. Create the colour image by using the Pixelmath expression (uncheck ‘Use a single RGB/K expression’, set Color space to ‘RGB color’):
    • R: ha
    • G: ho*ha+~ho*oiii
    • B: oiii
13. Apply CurvesTransformation. Start with saturation, proceed to individual channels.
14. Apply NoiseExterminator with default values or Denoise ~0.9, Detail ~0.55.
15. Stretch ‘ha_stars’ and ‘oiii_stars’ using HistogramTransformation. Use checkmark to track the histogram and use live preview to monitor the stretch amount. Drag middle slider to the left.
16. Apply the Foraxx process (step 10) to ‘ha_stars’ and ‘oiii_stars’.
17. Use CurvesTransformation to boost saturation of ‘foraxx_stars’.
18. Put the stars back using the Pixelmath expression: ~(~foraxx*~foraxx_stars).

And this is it! Here are some examples of my application of this process applied to various emission nebulae.

This emission nebula is one of the youngest star-forming regions in our galaxy. Some components of it are only few million years old. In the cosmic time scale, this is basically star birth happening in front of our eyes. This nebula also contains the hottest star found within 1 kpc of our Sun, the BD+66 1673, which has the surface temperature of 45,000 K and the luminosity 100,000 times that of the Sun. It is primarily responsible for ionizing the gas of the nebula and for compressing it by the strong stellar wind, leading to creation of new stars. 

This light travelled for 3,000 years before reaching my yard in Victoria, BC in June 2024.

NGC 7822 is a large target, filling the full-frame sensor of my camera attached to a 478 mm-long telescope. This is a 3.5 hr-long RGB exposure at f/5.9, processed using a Foraxx palette.

I captured this image of the Croc’s Eye Galaxy (M94, also called the Cat’s Eye Galaxy) yesterday in my yard in Victoria, BC, using a total of 3 hours of exposure. It’s an unusual galaxy – it has a an inner ring with a diameter of 5,400 light-years and an outer one with a diameter of 45,000 light-years. Pressure from the galactic core compresses the gas and dust clouds in the outer ring, where gravity pulls them together to form new stars. These stars pull in more gas and dust, resulting in a relatively empty region separating them from yet another layer of gas at the periphery of the galaxy.

M94 has a remarkably low amount of dark matter for a galaxy – the stars comprise almost all of its mass. Their light travelled for 16 million years before reaching my telescope.