I thought it might be interesting to find out how much time I invest to create an image like that, so I started writing down the whole process (at the end a bit fuzzy, but I believe everybody knows what fine tuning can be done). Here the whole story, 21.3 hours ... oh my, did not expect THIS much time!

1: Obtained over 3 nights the Ha, OIII and SII images (setup time for each session since I do not have a permanent observatory: 20 minutes, total 1 hour). Every filter required focus with Bahtinov mask (used Vega). Start time: ~1:10am. Let the session run for 4 hours (timer) so that the scope does not hit the mount. Did not want to get up after 4 hours again to stop it and it was anyway already close to dawn (~12 hours total integration time).

2: Copied from "field" laptop the images to a USB drive (15 minutes in total).

3: Copied images from the USB drive to the processing drive of the desktop (15 minutes in total).

4: Blinked all imges and filtered out bad ones. I was lucky with this set, no satellites, but I had to eliminate 15 SII images due to clouds. (20 minutes).

5: Created flat files in the morning for Ha and OIII (SII I was not able to find time yet). Copied files the same way as the lights (1 hour total time, including setup).

6. If you do not have already dark files you need to create them. This can take up to 10 hours in the night (for different exposure times). Temperature has to match!! My inages were taken at -18 degrees, luckily I had already a good set for that.

7: Started up Deep Sky Stacker (DSS) and selected all images, used Ha flat files and already obtained dark files. Setup for light, dark and flat all median, 1x Drizzle, Best 90%. Let it stack (45 minutes on Pentium 9900 processor, 16 GB RAM).

8: Found out that the Ha flats were not good, so used OIII flats instead. Did not re-register files, just stack. (30 mintes)

9: OIII flats are not perfect either, but better. Will decide to combine these 2 images in PixelMath to one (50% each).

10. Generated now the different channels (Ha, OIII, SII). Used "best" image determined by DSS as reference file, same parameter as "Lum" file, best 100%. Each roughly 10 minutes of work. (30 minutes)

11. Renamed Autosave files accordingly (NGC6823-L-OIIIFlat/LHa-Flat/Ha/OIII/SII) and moved them into the processing directory of NGC 6823.

12. As I thought, used PixelMath and added the 2 Lum files with "max(NGC6823_L_HaFlat_DSS_01,NGC6823_L_OIIIFlat_DSS_01)". Renemed resulting file as Lum file and saved as .xisf. (1 minute)

13. PixInsight time, DSS is done. If I am crazy I will do the same DSS tasks with 3x Drizzle and Sigma-Kappa clipping (2.0), will take much longer (up to 10 hours). Since this is a nebula I do not think it is necessary.

14. Dynamic Crop, just the edges due to stacking and not perfect alignment of each subframe. Mount did well though (and Plate Solving), not much to cut.

15. Created Dynamic PSF, Star Mask for local Deringing and Star Mask on STFed image, created previews to find the best global dark and bright setting in the Deconvolution, applied on preview the best setting with 100 iterations and confirmed improvement, applied found settings to whole image. Done with all 4 linear base images. Per image 20 minutes (80 minutes total).

16. DBE: selected the empty spots in the LUM image, saved the final result and applied with this set DBE to all 4 base images. (15 minutes, selecting takes some time)

17. Linear Fit: picked "brightest" base image (in this case the OIII image) and applied linear fit to all other base images (5 minutes)

18. Created RGB (in this case SHO) and LRGB (LSHO) images

19. Applied Gary's instruction how to use TGV (10 minutes) on the SHO image, that worked out great Gary (do not think that I understood everything, but most of it, will have to play around a bit more with that)!!

20. Applied MTT to SHO image (5 minutes).

21. Applied TGV and MTT to Lum image (5 minutes).

22. Stretched the SHO image, first slightly with Arsinh stretch, the HT.

23. Applied Lum to the SHO image.

24. Applied the corrections to get the nice Hubble palette image and found out that it did not come out nice, not at all (10 minutes). This is in generala problem with the SHO images, very good data are needed to get a clean SHO (adjusted). Might do something wrong here.

25. Created an HSO image instead, and did the same adjustments done to the SHO image (10 minutes).

26. HSO image looked much nicer, and now the tiny adjustments game starts. Flipped back and forth between PI and PS, each toolk has some nice features, some work, some do not. In total I spent approximately 2 hours to fine tune, and this can go much longer. For now this is it though (2 hours).

1276 minutes spent in total = 21.3 hours for an image that is for me quite nice, but does not compare to others here on Astrobin. So beginner, be warned, you have to spend a lot of time (and money, I would need some equipment upgrade) to get an image that looks nice enough in your eyes! And I am pretty sure that I did here and there some crucial mistakes. One part that I did not like at all: my base images had streaks in them, not sure where they do come from. I noticed them before, but not as dramatic as in this set. Might be light pollution from the road nearby, since I started taking the images lower at the horizon in the northeast direction, where it is the worst.