M47 (Open star cluster) & M42 (Final revision)

[RickBG]

M47 is an open cluster taken on Feb. 9, 2021.
CPC 8" SCT/ZWO 183 color/HyperStar 1.9 focal reducer/Optolong L-Pro filter
SharpCap Pro 3.3/PhotoShop
Gain: 115/Exposure 2.85"/Raw 16/Stacked frames 1,128/Capture Area 5496x3672 then cropped


Orion Nebula revised from the prior acquisition. Applied slight Color change and further stretching to bring out more detail.

[oopfan]

Hi Rick,

I love star clusters, but I've been searching for answers as to why your M47 looks the way it does. Reasons that come to mind are:
1. You set the black level too high, therefore cutting off the light from many faint stars.
2. You set your camera's bias/offset level too low.

Can you share the CameraSettings file for your light frames?

Some thoughts:
1. Star clusters are notoriously high dynamic range objects, therefore it is best to run it at your lowest gain setting. Your camera has a maximum full well depth of 15,000 electrons at the lowest gain. In your image, you chose gain 115 which lowered your full well depth to 4,000 electrons, the result of which is bloated, saturated stars.
2. From outward appearances, your light pollution is high, since I see that you used an Optolong L-Pro filter and an exposure of 2.85 seconds. So, assuming that your LP is high, I know that your highest noise source is light pollution. By comparison, your Read Noise is small at 2.2 electrons. If you were to choose the lowest gain, then the Read Noise increases to 3.0 electrons, a nearly imperceptible difference. However, the benefit of lowering the gain is that it raises the full well depth from 4,000 to 15,000 electrons.
3. I downloaded your image to my desktop and zoomed in. The faintest star I can find is magnitude 15.6, and a nearby saturated, bright star is magnitude 8.8. This results in a luminosity ratio of 10^((15.6 - 8.8) / 2.5) = 525. Therefore, the dynamic range is log(525) / log(2) = 9 stops. The ZWO documentation says that the dynamic range at gain 115 is 11 stops. So, somewhere you are losing 2 stops. My theory, as I said earlier, is that the bias/offset is too low, or you raised the black level in Photoshop or elsewhere.

I am looking forward to helping solve this problem since you invested a lot of money in your kit. I think that with some tweaking here and there that you can get more bang for your buck.

Brian

[RickBG]

Hi Brian,
Thanks for the reply and do appreciate your input. Above is a redo of the same data taken at the same time but with different processing. I increased the image size to 300% to blend in the pixels in PhotoShop and the Black point was adjusted to a low of 4 for RGB but in the first image the black level was very high as you have alluded to as a possible cause of your concern. In the first image, I used Astronomy Tools to intentionally make the stars fuzzy which increased the stars size but dimmed it as well whereas some might have used "reduce star size", based on their preferences. In SharpCap the Gain was automatically set to 115 by the Brain when 16 bits was chosen. Let me know if this image resolves your concern. Below are the data you requested and then some:
Rick
Debayer Preview=On
Pan=0
Tilt=0In
Output Format=FITS files (*.fits)
Binning=1
Capture Area=5496x3672
Colour Space=RAW16
Hardware Binning=Off
High Speed Mode=On
Turbo USB=100(Auto)
Flip=Horiz
Frame Rate Limit=Maximum
Gain=115
Exposure=2.855107
Timestamp Frames=Off
White Bal (B)=53
White Bal (R)=82
Brightness=4
Temperature=25
Cooler Power=100
Target Temperature=0
Cooler=On
Auto Exp Max Gain=326
Auto Exp Max Exp M S=30000
Auto Exp Target Brightness=158
Mono Bin=Off
Banding Threshold=35
Banding Suppression=0
Apply Flat=None
Subtract Dark=None
#Black Point
Display Black Point=0
#MidTone Point
Display MidTone Point=0.5
#White Point
Display White Point=1
Notes=
TimeStamp=2021-02-10T03:39:20.2835774Z
SharpCapVersion=3.3.7253.0
TotalExposure(s)=3220.56069599996
StackedFrames=1128
LiveStack.SaveRawFrames=None
LiveStack.AlignFrames=True
LiveStack.Derotate=True
LiveStack.StarsForAlignment=15
LiveStack.MinStarSize=3
LiveStack.MaxStarWidth=16
LiveStack.AlignAutoDisabled=False
LiveStack.ReduceNoiseAmount=0.5
LiveStack.DigitalGain=Off
LiveStack.BlackThreshold=20
LiveStack.BrightnessLimit=79.72973
LiveStack.FilterBrightness=True
LiveStack.AutoBrightnessLimit=False
LiveStack.FWHMLimit=7.216213
LiveStack.FilterFWHM=True
LiveStack.WarningThreshold=5
LiveStack.StackingMode=0
LiveStack.SigmaClipInitialCount=15
LiveSack.SigmaClipThreshold=3.492064
LiveStack.SigmaClipLowLimit=5.026454
LiveStack.AutoSaveReset=False
LiveStack.AutoSaveResetMinutes=30
LiveStack.LogarithmicHistogramHorizontalAxis=False
LiveStack.AutoSave=False
LiveStack.BlackLevel=0
LiveStack.WhiteLevel=100
LiveStack.MidLevel=25
LiveStack.WBB.db=0
LiveStack.WBG.db=0
LiveStack.WBR.db=0
LiveStack.Saturation=1.252567
AutoStretch.Strength=2.01005

[timh]

Hi Rick,

Like Brian I think the star cluster image does look rather flat - and the extra processing to have removed the colours?

In the past I have obtained images that look very similar to your M47 and it is has usually been under conditions where the background sky was just too bright and the object low in the sky? Under such circumstances there is inherently too little contrast and the dynamic range gets used up. --- and for something like a star cluster your total exposure time of 50 min at F 1.9 does seems like an awful a lot of light.

To help put expectations into context .... my F 4.5 system (and Bortle 6 skies) starts to deliver quite good images of relatively very bright objects like star clusters and M42 even within less than a minute - i.e . typical conditions for semi-live EEA - and with the appropriate dark/ flat applied - and just looking at the developing live stack in SC. Your system is more than 4 X faster than that even.

Anyway don't know if any of this helps at all Just thought it might well be worth looking at the early live stack as it develops to see if the image after 30s or so is really much different from that at 30 min?

Tim

[oopfan]

Hi Rick,

I created a side-by-side comparison of your original image and the reworked one. I zoomed in:

Rick M47 before and after 1.png (104.47 KiB) Viewed 1399 times

I annotated the image to show, what I believe to be, artifacts of aggressive denoising. Can you have another go at it? But this time:
1. No denoising.
2. Show outer space as a shade of gray. Do this by:
a. Easing off on contrast.
b. Lower the black point.

Last night I messaged my cohort who has a 183-based camera, but his is an Altair 183C, whereas yours is a ZWO ASI183MC. I asked if he thought that "Brightness=4" was too low. He said that he couldn't tell because every manufacturer of 183 cameras call it something different (e.g. Brightness, Offset, Bias). I would have a better idea if I knew the maximum value of the control. For example, if it is "4 out of 5" or "4 out of 10" then I think you are OK. But if it is "4 out of 20" then you have a problem.

I've attached a screenshot of a bias frame that shows what a properly adjusted brightness level looks like. You want to see both tails of the histogram as shown. If the brightness level is too low then the histogram shifts to the left and you start losing data. Do you have any bias frames laying around? The program "FITS Liberator" is free; just do an internet search.

Fortunately you can capture bias frames in the comfort of your home. Simply cover the objective, the same way as if you were creating darks. (I'm not sure how to do this with Hyperstar getting in the way.) Then, set the exposure to 1 millisecond and let her rip. Set gain to 115 and brightness 4, then capture several frames. To be thorough, capture some more frames at a few different brightness levels. Open them in FITS Liberator, take a screenshot, and share them with us here.

sample bias frame.png (161.99 KiB) Viewed 1399 times

Thanks,
Brian

[oopfan]

Hi Rick,

I searched for images on AstroBin that used the ASI183MC. I found this:
https://www.astrobin.com/409604/0/

He used SharpCap 3.2 and I assume LiveStack since he didn't mention any other software.

He's got a Celestron C8 SCT like you. He doesn't have HyperStar, but he is using a 0.4x focal reducer. That puts him at f/4 whereas you are running at f/2. Also, he used an Optolong L-Pro like you, so he does have light pollution. His kit seems fairly close to yours.

The stars have good color. By my estimation he is seeing down to about magnitude 18 with only 2.5 minutes integration time.

May I ask, what is your Bortle class? If you don't know, can you give me your approximate city/state. I can look it up.

Brian

[RickBG]

Hi Brian,

This is my third iteration which now shows more depth and color. I use levels, curves and saturation in PhotoShop. No noise reduction nor sharpening was used. The image is slightly cropped and is similar to the M13 you attached to your post. I find when the image is overly cropped the stars become bloated and fuzzy. If you crop (or zoom in) to M13 the stars start to look that way. This is my first star cluster image taken, therefore I believe it would be helpful, if you can show me one of your M47 images to help me improve on mine, even though LP and a different setup was used. I believe I'm slightly out of focus as I zoom into the stars and therefore I must check my focus every time I setup before acquisition. Thanks Brian for your persistence and effort to solve this enigma...I do appreciate it. With that persistence you showed, it encouraged me to improve on the image.

Rick

[oopfan]

Hi Rick,

I haven't imaged M47 but I can show you my Perseus Double Cluster from a couple months ago:

Perseus-Double-Cluster_W12-26x60s_R-70x60s_G-35x60s_APP_S15-3-25_SA-35-25_G94_SC_AFF_Denoise-2.jpg (598.47 KiB) Viewed 1361 times

I was rushing to finish it before it passed behind trees so I had poor focus in two of the three filters. But you see how my sky background is a shade of dark gray? I do that to show the limits of detection.

I like your new image better without the severe crop from before, but it is still too dark. There is still a lot more information in there that I'd like to see. Your contrast is still too high. In Photoshop, lower the contrast and increase the brightness. What does the image look like if you don't adjust 'curves'?

I can see a couple red and blue squiggles. Those are hot pixels. Did you capture darks? There is also a mysterious greenish glow near the center of the image. Somehow I remember someone on the forum having a similar artifact. Many of us thought it was dew, but I don't remember if it was ever resolved.

I am still not convinced that the "brightness" is set properly. I am not talking about Photoshop's brightness. I am referring to a setting on the camera that you make with SharpCap when you capture the images. This could explain why your image is dark. Several years ago when I started in astrophotograpjy, I didn't know about "brightness". I captured a star cluster called NGC 188. When I processed it, the sky background looked like "The Matrix" movie. I would like you to show me "The Matrix" in your image, if it exists. The way to do that is to cut the contrast in Photoshop.

Thanks,
Brian

[RickBG]

Hi Brian,

This is now my fourth iteration, all coming from the same initial data. By increasing the number of levels and curves, more and more stars start to appear and turn brighter. It gets to point when enough is enough. Some feel too many stars takes away from the main star cluster or nebulae. With each iteration, I find the image is starting to look better and more pleasing to the eye. I removed the line artifacts you mentioned with the "eraser" in PhotoShop which takes seconds. Sometimes it is difficult to know when we are "full or still hungry" and often times we over do it. As I had pointed out, when the image is overly cropped and processed, the stars appear too bloated and fuzzy like cotton balls. This is what I see in the image you presented. There are some who find it pleasing and there are those who feel the stars ought to be round without spikes (barring the fact that Newtonians will cause spikes). Others feel the night sky should look dark as dark can be, as if you are out in the countryside with zero light pollution. I encourage you to zoom into your image and note the red lines that are superimposed over the star spikes...I don't believe they should be there. There is also excessive noise in the image which suggest over processing. With less processing your stars will look more natural and less bloated and spikey. I genuinely feel that processing is an art and not a science and a lot of it has to do with your taste. I decided to look up M47 on line, and no two photos are alike as to be expected. They vary from one extreme to another; some I like and others I feel should have never been posted in the first place. Though our tastes are different, through your persistence, I did 4 iterations which I'm grateful for...because there is always room for improvement and I believe it shows. Thanks!

Rick

[turfpit]

Rick

Have a look at:

viewtopic.php?f=23&t=3694&p=19530#p19530

viewtopic.php?t=596

Note the timescales - it might give you some encouragement.

Dave