Astro Pixel Processor Help Wanted :-) - Forums

Topic starter

Feb 8, 2021

Hello, hope you are all doing well.
I am addressing experienced users of Astro Pixel Processor:
This is a neutralized 15% BG stretched unprocessed image. When removing light pollution, where should I draw 5 or more boxes and how large or small should the boxes be? My boxes make everything look worse (still a trial user).

Thanks and Cheers
Robert

Nick Axaris

Feb 8, 2021 · Edited: Feb 8, 2021

Smaller boxes and more of them evenly spread except for the nebula areas, you need to experiment around a bit to get it right

Nick Axaris

Feb 8, 2021

Also do you have the before image?

astropical

Topic starter

Feb 8, 2021 · Edited: Feb 8, 2021

Nick Axaris:
Also do you have the before image?

Hello Nick, thank you,
This is the before image.

Bradley Watson

Feb 8, 2021

Hi Robert, I would also crop your image to get rid of the stacking artifacts around the edge, if included in your selection it will add to the LP model.

Like said before small boxes. The light pollution function also calibrates the background in the newer version so include dark areas also as as well as obvious LP.

Avoid nubula, I have heard this is not necessary but I avoid stars also. Make sure you get the edges of your image too, but play around

Good luck and CS
Brad

Tim Richter

Feb 8, 2021

Hello Robert,

i think i can help you. My approach is to place a lot of small sample boxes. Very important:

Don't include stars, don't include nebula regions, don't include anything belonging to the object. Only include background sky. To get an idea of how small i make these boxes: In my recent

Andromeda Galaxy
image, there were lot's of stars to avoid. APP was telling me quite often that i should make the boxes bigger than the 32x32 minimum. Try to spread them evenly throughout the entire image, whenever possible. I recently caught myself placing 150 boxes in an image, that's probalby overkill.

Another tip on how to improve the placing: Turn the preview to maximum, and enable saturation. Since APP has a threshold for saturation, you will see which regions belong to nebulae, or if they are background sky. After applying, i often zoom in and out to place some more, if i think that e.g. the corner is not evenly neutralized.

I hope i could help!

May the night be with us!

Helpful Supportive

astropical

Topic starter

Feb 8, 2021

Bradley Watson:
Hi Robert, I would also crop your image to get rid of the stacking artifacts around the edge, if included in your selection it will add to the LP model.

Like said before small boxes. The light pollution function also calibrates the background in the newer version so include dark areas also as as well as obvious LP.

Avoid nubula, I have heard this is not necessary but I avoid stars also. Make sure you get the edges of your image too, but play around

Good luck and CS
Brad

Thanks so much Bradley, crop is a good idea. Why did I not think about it? :-)
You gave me a good vision of what to do.
Cheers and CS
Robert

astropical

Topic starter

Feb 8, 2021

Tim Richter:
Hello Robert,

i think i can help you. My approach is to place a lot of small sample boxes. Very important:

Don't include stars, don't include nebula regions, don't include anything belonging to the object. Only include background sky. To get an idea of how small i make these boxes: In my recent

Andromeda Galaxy
image, there were lot's of stars to avoid. APP was telling me quite often that i should make the boxes bigger than the 32x32 minimum. Try to spread them evenly throughout the entire image, whenever possible. I recently caught myself placing 150 boxes in an image, that's probalby overkill.

Another tip on how to improve the placing: Turn the preview to maximum, and enable saturation. Since APP has a threshold for saturation, you will see which regions belong to nebulae, or if they are background sky. After applying, i often zoom in and out to place some more, if i think that e.g. the corner is not evenly neutralized.

I hope i could help!

May the night be with us!

Hello Tim,
You did help a lot indeed, thanks a ton.
I may need the whole night but I will try :-)
Cheers and CS
Robert

Andy 01

IOTD Staff

Feb 8, 2021

Looks like you need better flats. How many did you do & are they even & consistant?

astropical

Topic starter

Feb 8, 2021 · Edited: Feb 8, 2021

Andy 01:
Looks like you need better flats. How many did you do & are they even & consistant?

Thanks a lot Andy, I reused 20x formerly taken flats but they were obsolete 🥴
It was a pretty humid night possibly with thin cloud or moisture layers high above.
Integration was 2.4 hours, yet only faint nebulosity captured.
Cheers
Robert

astropical

Topic starter

Feb 8, 2021

@Bradley Watson
@Tim Richter
Gentlemen, I followed your game rules and obtained a good result after drawing about 200 boxes, then removing the red and yellow ones and all over again. Then tuning in Photoshop. Oh boy, all that does take time. 😆 Reckon this image is kinda worst case.

The best solution is to avoid light. It was a pretty humid night possibly with thin cloud or moisture layers high above.
In spite of 2.4 hours integration, only faint nebulosity was captured and it required the help of Starnet++ to bring it up.

Anyway, I am getting somewhere with APP thanks to you. Problem solved.
CheerS
Robert

Bradley Watson

Feb 8, 2021

Very nice job @astropical. I am glad the advice helped. Part of the reason I use APP is because of the LP removal tool, its a very good one once you get the hang of it.
CS
Brad

Gabriel R. Santos (grsotnas)

Feb 8, 2021

@astropical

Your biggest problem is probably in the Calibration phase: my bet is your flats did not work correctly, which generated the gradients, which are much harder to remove afterwards, as your image shows. It could also be a blazingly bright street light near the object, or significant light contamination/internal reflections. But my bet is still in the Calibration ;-)

In your case, I can only think on the approach alreade mentioned here: a dense sample grid. Think 50 to 100 samples. But bear in mind that by doing so you end up killing background/faint nebulosity and background "richness". It works reasonably well on isolated objects (think isolated galaxies), but it is always more harmful to Milky Way fields (think Orion nebula or Sagittarius region).

I would advise you to really look into learning the Calibration part well. When it works correctly, you are usually left with a smooth light pollution gradient that is straightforward to remove with a few 5-10 LPC samples.

Helpful Insightful Respectful Engaging Supportive

astropical

Topic starter

Feb 8, 2021

Thanks so much Gabriel,
Actually, this image is not calibrated at all, a worst case captured through humid air possibly a thin cloud layer as well.
Absolutely, I agree, it will be critical with the Milky Way in the background, probably worse than light :-)
Your heads up about flats is much appreciated. I should not be lazy and take new flats after every session.
Please stay safe!
Cheers
Robert

1074j

Feb 8, 2021

Bradley Watson:
Hi Robert, I would also crop your image to get rid of the stacking artifacts around the edge, if included in your selection it will add to the LP model.

Like said before small boxes. The light pollution function also calibrates the background in the newer version so include dark areas also as as well as obvious LP.

Avoid nubula, I have heard this is not necessary but I avoid stars also. Make sure you get the edges of your image too, but play around

Good luck and CS
Brad

Thanks so much Bradley, crop is a good idea. Why did I not think about it? :-)
You gave me a good vision of what to do.
Cheers and CS
Robert

I second the cropping advice. Very important.