Monday, June 16th was exactly 1 year since my first astrophotography session with a 85mm lens on an astro-modified Nikon Z5 MILC and a SW SA GTi mount. Since then, I have tried various camera lenses and 3 smallish refractors. My current rig is a WO GT81 IV + Flat GT with a ZWO ASI585MC Pro on a ZWO AM3. I use a ZWO ASIAir Mini for capture/control. Here is my latest astrophotograph with this rig:
Ha+RGB+IR:
https://app.astrobin.com/u/Astro%C3%85mazer?i=i8ask6#gallery
RGB+IR:
https://app.astrobin.com/u/Astro%C3%85mazer?i=i8ask6&r=B
This is going to be a very detailed post so thanks in advance for reading till the end!
Acquisition wise, here is my most recent process:
1. Try to plan Luminance around 3rd quarter to 1st quarter and Narrow Band around the rest of the month.
2. Try to stay at least 75 degrees away from the moon even when using NB filters
3. If seeing is bad, or have to get Luminance data between 1st quarter and 3rd quarter of the moon, use the IR Pass filter
4. Get at least 16 darks, 32 flats, 16 dark flats (Re-thinking this, may not need darks)
5. If initial polar alignment is off by more than 2 degrees, repeat after 1st try
6. If changes to RA affects Dec while polar alignment, mount isn't perfectly level, bump up guiding aggression a little
7. Re-run guiding calibration if rig was disassembled/moved
8. Acclimatize scope ~30mins-1hr before session
9. Autofocus every 2hrs/2 degrees C change in ambient temperature
10. Dither, dither, dither - though lost ~40 pixels each side on this project before field rotation issues between nights.
10. Gain 252, 30s subs for stars, 60s for Lum, 90s for IR Pass, 180s for NB
11. There is some tilt in my imaging train which I am trying to fix!!!
I am using a WO 120mm guide scope with a ASI220MM Mini and getting ~0.6" guiding accuracy. Mean FWHM has been between 3"-4" recently. On a bad night like June 3rd, the sky looks very orange (smoke from Canadian wildfires?). On a good night like June 4th, I can see Andromeda rising with my 10x42mm binoculars. Have had less than 10 clear nights all spring.
Processing wise, I use Siril/Gimp/GraXpert/StarNet++. Here is what I have been doing:
1. I use a modified version of the OSC preprocessing script with drizzle, in Siril. Modifications:
a. Use weighted FWHM as a metric for stacking.
b. When stacking Ha, stack on red channel.
2. Stack data from each night separately.
a. I can write a script to customize this so that all calibration is done with flats etc per night and registration/stacking is done with all subs at once.
b. Do not have hard disk space to do all at once registration/stacking sometimes so do my own weighted stacking of different batches.
3. I keep the registration FWHM and background CSVs and use them to weigh each night's stack.
4. Once I have the stacks for each filter, I do the following in linear:
a. Desaturate stars if needed (does this make my data non-linear?)
b. Run the AberrationRemover.py script.
c. GraXpert background extraction.
d. Plate solve.
e. Color calibration - Spectrophotometric if filter analogue is available else Photometric.
5. At this point I remove the stars. For the stars, I use the Modified ArchSinH Transform. Stretch to my liking, add some saturation etc.
6. As for stretching the starless image, I process each filter separately and blend in the stretched data. Try to keep the sky brightness peak < 0.1 and bring out the features I am after.
7. I mostly use the StatisticalStretch.py script, GHS and some Curves, maybe some CLAHE at the end.
a. In GHS, I use Independent Channel Values or Saturation for the stretch model and RGB blend for the clip mode.
b. Have noticed other combinations can lead to values > 1
8. Then I go over to Gimp, reduce the color noise in the background with the Hue-Chroma tool. Add some cosmetic healing if the stars were not removed well enough or have issues with flats.
a. I have tried decomposing to Lab space and sharpening L, blurring a and b to some level of success but mostly reducing chroma and lightness a little in the Hue-Chroma tool is good enough.
9. After blending in the broadband data to my liking (In this example, I weighed the Lum and IR data by FWHM and background level) I get Continuum Subtracted NB using the Siril script and use Pixel Math to compose my final image.
10. If I am doing a SHO processing, I try Dynamic NB equations or custom static blends based on the IMX585 sensor's response to the NB wavelengths.
11. Finally, blend in the stars using Pixel Math. Apply subtle GHS/curves/saturation adjustments if required.
12. Run GraXpert denoise till the grain is smoothened out. I feel like I am denoising too much. But if I don't, I end up with a lot of speckle/grain.
Again, thanks if you stuck around till here. And thanks in advance for any feedback!!!
CS
GD
Ha+RGB+IR:
https://app.astrobin.com/u/Astro%C3%85mazer?i=i8ask6#gallery
RGB+IR:
https://app.astrobin.com/u/Astro%C3%85mazer?i=i8ask6&r=B
This is going to be a very detailed post so thanks in advance for reading till the end!
Acquisition wise, here is my most recent process:
1. Try to plan Luminance around 3rd quarter to 1st quarter and Narrow Band around the rest of the month.
2. Try to stay at least 75 degrees away from the moon even when using NB filters
3. If seeing is bad, or have to get Luminance data between 1st quarter and 3rd quarter of the moon, use the IR Pass filter
4. Get at least 16 darks, 32 flats, 16 dark flats (Re-thinking this, may not need darks)
5. If initial polar alignment is off by more than 2 degrees, repeat after 1st try
6. If changes to RA affects Dec while polar alignment, mount isn't perfectly level, bump up guiding aggression a little
7. Re-run guiding calibration if rig was disassembled/moved
8. Acclimatize scope ~30mins-1hr before session
9. Autofocus every 2hrs/2 degrees C change in ambient temperature
10. Dither, dither, dither - though lost ~40 pixels each side on this project before field rotation issues between nights.
10. Gain 252, 30s subs for stars, 60s for Lum, 90s for IR Pass, 180s for NB
11. There is some tilt in my imaging train which I am trying to fix!!!
I am using a WO 120mm guide scope with a ASI220MM Mini and getting ~0.6" guiding accuracy. Mean FWHM has been between 3"-4" recently. On a bad night like June 3rd, the sky looks very orange (smoke from Canadian wildfires?). On a good night like June 4th, I can see Andromeda rising with my 10x42mm binoculars. Have had less than 10 clear nights all spring.
Processing wise, I use Siril/Gimp/GraXpert/StarNet++. Here is what I have been doing:
1. I use a modified version of the OSC preprocessing script with drizzle, in Siril. Modifications:
a. Use weighted FWHM as a metric for stacking.
b. When stacking Ha, stack on red channel.
2. Stack data from each night separately.
a. I can write a script to customize this so that all calibration is done with flats etc per night and registration/stacking is done with all subs at once.
b. Do not have hard disk space to do all at once registration/stacking sometimes so do my own weighted stacking of different batches.
3. I keep the registration FWHM and background CSVs and use them to weigh each night's stack.
4. Once I have the stacks for each filter, I do the following in linear:
a. Desaturate stars if needed (does this make my data non-linear?)
b. Run the AberrationRemover.py script.
c. GraXpert background extraction.
d. Plate solve.
e. Color calibration - Spectrophotometric if filter analogue is available else Photometric.
5. At this point I remove the stars. For the stars, I use the Modified ArchSinH Transform. Stretch to my liking, add some saturation etc.
6. As for stretching the starless image, I process each filter separately and blend in the stretched data. Try to keep the sky brightness peak < 0.1 and bring out the features I am after.
7. I mostly use the StatisticalStretch.py script, GHS and some Curves, maybe some CLAHE at the end.
a. In GHS, I use Independent Channel Values or Saturation for the stretch model and RGB blend for the clip mode.
b. Have noticed other combinations can lead to values > 1
8. Then I go over to Gimp, reduce the color noise in the background with the Hue-Chroma tool. Add some cosmetic healing if the stars were not removed well enough or have issues with flats.
a. I have tried decomposing to Lab space and sharpening L, blurring a and b to some level of success but mostly reducing chroma and lightness a little in the Hue-Chroma tool is good enough.
9. After blending in the broadband data to my liking (In this example, I weighed the Lum and IR data by FWHM and background level) I get Continuum Subtracted NB using the Siril script and use Pixel Math to compose my final image.
10. If I am doing a SHO processing, I try Dynamic NB equations or custom static blends based on the IMX585 sensor's response to the NB wavelengths.
11. Finally, blend in the stars using Pixel Math. Apply subtle GHS/curves/saturation adjustments if required.
12. Run GraXpert denoise till the grain is smoothened out. I feel like I am denoising too much. But if I don't, I end up with a lot of speckle/grain.
Again, thanks if you stuck around till here. And thanks in advance for any feedback!!!
CS
GD
