[RCC] advice required on processing of nebula images

I am a beginner to AP and I think I am doing quite well with acquisition now, but I am fairly unsure of how to go about processing. I would really appreciate any help or feedback on how you think I have done on my latest effort.

This is a little over 6h of data of The Cygnus Wall from an Esprit 150 ED (reduced to f/5.4) and a ASI2600 MC Pro. Unguided.

I stacked in Astropixel Processor using the default settings on the best 90% of subs and then applied the AstroFlatsPro tool in Photoshop.

I'm attaching a link to the FITS file too so you can play with it yourselves.

Thanks in advance.

FITS https://drive.google.com/file/d/1iVU46fEF7bSZLQaj5AIkGnyR4zaGQBLD/view?usp=sharing

Let's start by producing an unclipped image because the one you sharing is massively clipped in the hystogram (and colour uncalibrated). I do not use or like APP so I cannot advise you in this. I know what I'd do in PI or Siril but there is where is stops. Ah, don't use PS to flatten images.

andrea tasselli:
Let's start by producing an unclipped image because the one you sharing is massively clipped in the hystogram (and colour uncalibrated). I do not use or like APP so I cannot advise you in this. I know what I'd do in PI or Siril but there is where is stops. Ah, don't use PS to flatten images.

Thanks Andrea. The image I added to the post was stretched using APP, so here is the un-stretched image

Brian, it is all black. You should post the unstretched image in the same folder of before.

andrea tasselli:
Brian, it is all black. You should post the unstretched image in the same folder of before.

I am not sure what you mean? The unstretched image in APP looks black when unstretched. The colours only appear when the image has been stretched. This is what I posted.

Or do you mean the FITS file? I don't think that changes when you stretch the image in APP. I think the stretch only applies to the JPG version.

The histogram adjustments are the first thing to sort out. Watch a few tutorial videos on histograms on a site like YouTube and then get to work balancing the colors and fixing the clipping mentioned above. You should be able to do these adjustments in Photoshop, which it looks like you are already using, if you can't figure it out in Astropixel Processor.

Once you get the histogram fixed, one of the first things to work on is the color (chrominance) noise in the background. In the first image shown, zooming in on the darker regions shows a lot of green pixels popping up. If you download the free program Siril, you could run this image through the Remove Green Noise function, and it will make a dramatic improvement. Before doing so, you should have the colors properly balanced in the histogram, however.

The good news is that it looks like you have good data with which to work. Many people try to learn processing with sub-standard data that came from their first efforts with a telescope, which makes things much more difficult. (That would be me back in 2012!)

Stuart Taylor:
Or do you mean the FITS file? I don't think that changes when you stretch the image in APP. I think the stretch only applies to the JPG version.

Apparently, it does!

The histogram adjustments are the first thing to sort out. Watch a few tutorial videos on histograms on a site like YouTube and then get to work balancing the colors and fixing the clipping mentioned above. You should be able to do these adjustments in Photoshop, which it looks like you are already using, if you can't figure it out in Astropixel Processor.

Once you get the histogram fixed, one of the first things to work on is the color (chrominance) noise in the background. In the first image shown, zooming in on the darker regions shows a lot of green pixels popping up. If you download the free program Siril, you could run this image through the Remove Green Noise function, and it will make a dramatic improvement. Before doing so, you should have the colors properly balanced in the histogram, however.

The good news is that it looks like you have good data with which to work. Many people try to learn processing with sub-standard data that came from their first efforts with a telescope, which makes things much more difficult. (That would be me back in 2012!)

Thanks. In fact the stretching is done in APP using one of a few provided presets. I just select the one that looks best. But I shall try Siril. Thank you

andrea tasselli:
Stuart Taylor:
Or do you mean the FITS file? I don't think that changes when you stretch the image in APP. I think the stretch only applies to the JPG version.

Apparently, it does!

No, it doesn't. I asked this specific question in the APP support forum

Then your original, un-stretched image is significantly clipped in the histogram, which is even worse news.

As per the histogram on top overlayed to the original image you posted:

Stuart Taylor:
…

andrea tasselli:
Stuart Taylor:
Or do you mean the FITS file? I don't think that changes when you stretch the image in APP. I think the stretch only applies to the JPG version.

Apparently, it does!

No, it doesn't. I asked this specific question in the APP support forum

To shed some light on this: APP tries not to overwrite files but proposes a suffix to the file name. Stretched data can of course also be saved as FITS (or tif, jpeg, png etc.).

andrea tasselli:
Then your original, un-stretched image is significantly clipped in the histogram, which is even worse news.

As per the histogram on top overlayed to the original image you posted:

Sorry, I am not sure I understand you. Are you saying that the FITS file I shared in gDrive has already had a stretch applied to it? As far as I am aware, all APP does is to calibrate, register, stack the original subs and calibration frames etc. This is apparent from the fact that when I posted the un-stretched version (above) it was completely black.

This is the appearance in APP un-stretched, As you can see, completely black

The image was un-stretched but it was clipped, as far as I can tell.

Your image lacks bit depth.

In PixInsight, by running a script which evaluates the effective dynamic range of an image, I get for your integration:

Calculating effective dynamic range...
new_three_nights_best_90_
Channel | Values | Bits |
--------|----------|-------|
0 | 28334 | 14.79 |
1 | 28744 | 14.81 |
2 | 22709 | 14.47 |

An 90 x 60 sec integration from my Canon 550D which is a 14-bit DSLR camera gives ~17.5 bits. Something is going wrong during your calibration. Six hours from a 16-bit camera should at least give 16 bits of depth. You are either clipping during calibration or you were going back and forth between raw and non-raw formats.

(EDIT: I know I am oversimplifying, but I can't believe that my poor old (2010) DSLR camera beats a modern CMOS astro camera.)

If your images are clipped, it's likely you did not set the offset value of ASI2600 properly. I don't know what acquisition software you use, but in the camera settings there should be somewhere you can fill in the offset value. For ASI2600 the default is 50. Setting it too low or without setting it could cause clipped images. That's the mistake I made when I started out with my ASI294MC pro. (If you use an ASI Air, then the setting is fixed at 50 and the issue would lie somewhere else.)

andrea tasselli:
The image was un-stretched but it was clipped, as far as I can tell.

Ah. Apologies, I didn't understand what you meant by clipped. I assumed this meant the same as stretched.

Die Launische Diva:
Your image lacks bit depth.

In PixInsight, by running a script which evaluates the effective dynamic range of an image, I get for your integration:

Calculating effective dynamic range...
new_three_nights_best_90_
Channel | Values | Bits |
--------|----------|-------|
0 | 28334 | 14.79 |
1 | 28744 | 14.81 |
2 | 22709 | 14.47 |

An 90 x 60 sec integration from my Canon 550D which is a 14-bit DSLR camera gives ~17.5 bits. Something is going wrong during your calibration. Six hours from a 16-bit camera should at least give 16 bits of depth. You are either clipping during calibration or you were going back and forth between raw and non-raw formats.

(EDIT: I know I am oversimplifying, but I can't believe that my poor old (2010) DSLR camera beats a modern CMOS astro camera.)

I am not shooting in RAW, I am shooting in FITS

Yungshih Lee:
If your images are clipped, it's likely you did not set the offset value of ASI2600 properly. I don't know what acquisition software you use, but in the camera settings there should be somewhere you can fill in the offset value. For ASI2600 the default is 50. Setting it too low or without setting it could cause clipped images. That's the mistake I made when I started out with my ASI294MC pro. (If you use an ASI Air, then the setting is fixed at 50 and the issue would lie somewhere else.)

I use NINA for acquisition and set the offset at 50

Stuart Taylor:
I am not shooting in RAW, I am shooting in FITS

I was afraid of any undesired file format conversions (FITS to TIFF or something like that).

Die Launische Diva:
Stuart Taylor:
I am not shooting in RAW, I am shooting in FITS

I was afraid of any undesired file format conversions (FITS to TIFF or something like that).

I only convert to JPG at the end of the APP process. All my subs are FITS and the stacked final image produced by APP is also FITS.

So what does 'clipped' mean in the above answers?

It means there are gaps in the histogram, as I tried to show in my previous post. In practice your file while screen-stretched in PI does not show a continuous variation of shades, from the darker to the lighter but some discrete number of shades. It may be that the file you posted ins't the same you were working with earlier on.

The image I posted in the first post is a JPEG version of the FITS file I shared a gDrive link to. But I have done some processing of it in photoshop. It was the processing I was asking about - i.e. did I do a good job? I assumed that was what this forum is for. I then also shared the original FITS file with no processing done (so you guys could see what you could do with it (as you are all much more experienced than I am)

But if you need to see a JPG of the unprocessed version, here it is (just with one of the standard stretched applied in APP).

Regardless of what happened to that particular file the first image posted was pretty good, so bravo!
The main source of criticism is the colour balance which seems pretty off to me. Specifically I can't see those greens which should really be teal or light blue. Star colours I can't judge as it is pretty hard to make them out but a first glance I'd say they are way too green as well.

andrea tasselli:
Regardless of what happened to that particular file the first image posted was pretty good, so bravo!
The main source of criticism is the colour balance which seems pretty off to me. Specifically I can't see those greens which should really be teal or light blue. Star colours I can't judge as it is pretty hard to make them out but a first glance I'd say they are way too green as well.

Great! Thanks

How would I correct the colour balance? The selective colour tool in photoshop?

Am I wrong in assuming you have PI? If so open it in PI, go to the menu Process>ColorCalibration and select PhotometricColorCalibration. Insert the data relevant to your case unter the tab/heading: Image Parameters and give it a go.

There is samething very similar in Siril and there must be something of that sort in APP too. No Photoshop.

You should be able to adjust Levels and Curves in Photoshop to get the color where you want it to be. I'm not a Photoshop user, but I'm sure it has these functions somewhere.

andrea tasselli:
Am I wrong in assuming you have PI? If so open it in PI, go to the menu Process>ColorCalibration and select PhotometricColorCalibration. Insert the data relevant to your case unter the tab/heading: Image Parameters and give it a go.

There is samething very similar in Siril and there must be something of that sort in APP too. No Photoshop.

Thanks. No, I do not use PI. However, I just discovered Siril and I like it very much! It's super fast at stacking and the photometric colour calibration is magical! The first stacked image usually looks very green, but the photometric colour calibration tool fixes all that beautifully in one go! Just great!!

Hi Stuart, I have played a bit with your original FITS file. I am also using APP. First, I have used crop and rotate under 9) Tools:

Then I have used 9) Tools - remove light pollution. Selecting just a few small areas in dark spots I got this result:

Then, I adjusted the contrast a bit (set to 0.05), leaving SAT.TH unchanged (you can play with these parameters to see what is best for your image):

Then in GIMP (not yet using PS), in Color-Curves, I adjusted the blackpoint (~15), and adjusted the curve to follow its original track:

At this stage to can, again in Colors-Curves adjust each color as you desire. Here is just an example with a slight increase in blue around the center of the curve (using multiple anchor points):

Hope this helps ...

All the best
CS
Manuel

Related discussions