copy/paste of the deets from when I originally posted it on the reddit last november:

So despite doing astrophotography for over 5 years I’ve never gotten an image of M31 that I’ve completely liked until now. My last real attempt at it was in 2019, and looking back has some obvious noise reduction and color gradient issues (tbf I shot it from downtown Athens, GA). I recently visited my future in laws, who happen to live in the middle of nowhere, so I brought along my telescope to take some deep space pics. I also incorporated some hydrogen-alpha data from a while ago to help boost the Hii regions (the red splotches) in the galaxies a bit. At first I wasn’t sure about the addition of the Ha data, so I ended up processing this image twice from scratch, one with Ha and one without. There were parts of each that I liked and disliked, so I blended the two finished pics together, and was very pleased with the results, which is the final image you see above. also here’s a starless version for the hell of it. Captured on November 17th, 2022, from a Bortle 4 zone (Ha data was captured in October 2020 from a Bortle 6 zone)

Places where I host my other images:

Instagram | Flickr

Equipment:

  • TPO 6" F/4 Imaging Newtonian

  • Orion Sirius EQ-G

  • ZWO ASI1600MM-Pro

  • Skywatcher Quattro Coma Corrector

  • ZWO EFW 8x1.25"/31mm

  • Astronomik LRGB+CLS Filters- 31mm

  • Astrodon 31mm Ha 5nm, Oiii 3nm, Sii 5nm

  • Agena 50mm Deluxe Straight-Through Guide Scope

  • ZWO ASI-290mc for guiding

  • Moonlite Autofocuser

Acquisition: 11 hours 14 minutes (Camera at Unity Gain(Ha) and half unity (LRGB), -15°C)

  • L- 46x120"

  • R- 20x120"

  • G- 18x120"

  • B- 18x120"

  • Ha -47x300" x 2 panels

  • Darks- 30

  • Flats- 30 per filter

Capture Software:

  • Captured using N.I.N.A. and PHD2 for guiding and dithering.

PixInsight Processing:

  • BatchPreProcessing

  • SubframeSelector

  • StarAlignment

  • Blink

  • ImageIntegration

  • DrizzleIntegration (2x, Var β=1.5)

Linear:

  • DynamicCrop

  • automaticBackgroundExtraction

  • EZ Decon

  • NoiseXTerminator

  • ArcsinhStrecth + HistogramTransformation to bring nonlinear

RGB Linear:

  • Channelcombination to combine monochrome images into RGB image

  • PhotometricColorCalibration

  • SCNR green

  • HSV Repair

Adding Ha:

I followed this tutorial which had great results on some prior HaLRGB galaxy pics:

http://www.arciereceleste.it/tutorial-pixinsight/cat-tutorial-eng/85-enhance-galaxy-ha-eng

Ha-Q * (Red-med (Red))

Q=0.15

  • PixelMath to combine Clean Ha

  • PixelMath to add Ha to RGB image ($T)

R= $T+B*(Ha_Clean - med(Ha_Clean))

G= $T

B= $T+B*0.2*(Ha_Clean - med(Ha_Clean))

B=1

Nonlinear Processing:

The RGB image without Ha was processed almost identically as above, minus the Ha addition steps. while nonlinear they were processed pretty similarly, so I’m just going to list out the general steps as they broadly apply to both images:

Nonlinear:

  • LRGBCombination with stretched luminance

  • Shitloads of CurveTransformations to adjust lightness, saturation, contrast, hues, etc.

  • Extract L --> LRGBCombination for chrominance noise reduction

  • color saturation to slightly desaturate the Ha regions

  • More curves

  • SCNR > Invert > SCNR > invert to remove some greens and magentas

  • LocalHistogramEqualization

Two rounds of this. one at size 16 for the finer ‘feathery’ details and one at size 500 for large scale structures

  • even more curves

  • NoiseXTerminator

  • EZ Star Reduction

  • noise generator to add noise back into star reduced areas

  • guess what baby more curves

PixelMath to combine the two fully processed images 50:50

  • Another round of EZ star reduction, NoiseXTerminator, and LRGBCombination chrominance noise reduction

  • HEY DID YOU KNOW I LIKE USING THE CURVESTRANSFORMATION PROCESS?

  • Resample to 60%

  • Annotation