NGC 6888 HOO - a dissapointment

[Rodd]

Very disappointed.  I do not understand it.  The Ha and OIII data are good - FWHM of 2.5 and 2.2 respectively.  The medians are a bit elevated due to my sky, but with 13 hours of data, I expected something a bit finer. Maybe I just need more data.  I don't think that will be the answer unless is A LOT more data.  I barely see a difference between a 90 sub Ha stack and a 40 sub Ha stack--so 180 sub stack may not be that much different.  The OIII envelope is nice, with a FWHM of 2.2, but its noisy and not that strong.  My biggest issue is the background.  The Ha stack has a pretty full background, but the HOO image has a noisy, lack luster background  that at close in scales cab be seen to carry copious blue/green.  And this data has been binned 2x2--so the integration is more like 26 hours!

Question:  Should I concentrate on OIII or Ha.  I have one more clear night , and I get a maximum of 5 hours per night.  Maybe I should just collect data until I am satisfied--40-50 hours additional.  Bicolor images often give me trouble--this target has always given me trouble.  13 hours binned 2x2 should be enough.  Maybe Something is not right in my processing.  I have tweaked and adjusted the image to death.  I must start over, but probably will end up in the same place.

 

 

 

 

Edited May 30, 2022 by Rodd

[vlaiv]

I think it's rather nice.

Have you tried starless processing?

Maybe that way you'll have more control over what is happening in the image.

Care to share the data ?

 

[tomato]

I think you are being hard on yourself, this was the IKI winning entry, 80 hrs of premium Ha/OIII/SII data…

[Rodd]

Yes. I will upload the fits files when i get back to my computer a bit later today.

I have never tried starless. Well, a good starless.  I can mask out stars, but not perfectly.  I tend not to like starless versions, but as a processing aide it might be useful

Edited May 30, 2022 by Rodd

[Rodd]

3 minutes ago, tomato said:

I think you are being hard on yourself, this was the IKI winning entry, 80 hrs of premium Ha/OIII/SII data…

I have never like the sho palette for this target.  Like Thors helmet and the veil, for some reason I like the HOO versions.   

[teoria_del_big_bang]

I think it is a great image. I got some good Ha and even some decent SII data earlier this year but my OIII was disappointing but you seem to have a lot of that feint outer shell.
I certainly would be more than happy with it 🙂 

Steve

Edited May 30, 2022 by teoria_del_big_bang

[Rodd]

2 minutes ago, teoria_del_big_bang said:

I think it is a great image. I got some good Ha and even some decent SII data earlier this year but my OIII was disappointing but you seem to have a lot of that feint outer shell which mine was missing.

I certainly would be more than happy with it 🙂 

Steve

Thanks Steve.  Tonight is supposed to be clear— after this clouds and rain for a week.  I think this is the last image with the C11Edge before I change to a refractor for nebula work.  So one more night to add to it to make it better.  Ha or OIII?  Maybe Ha. Not sure

[Budgie1]

46 minutes ago, Rodd said:

I have never tried starless. Well, a good starless.  I can mask out stars, but not perfectly.  I tend not to like starless versions, but as a processing aide it might be useful

I'm not sure what software you use for processing but StarNet2 and StarXTerminator are both very good. Although they both allow you to remove the stars from the image, the stars are not lost, both of these also create an image with just the stars, which allows you to work on the stars and the background image separately. When you've finished processing both of them, you can put them back together again.

Working on background (nebula or galaxy) and the stars separately means you have more control over the adjustments you make. It also makes is easier if you want to use SHO or HOO for the main target and add RGB stars.  

[gorann]

2 hours ago, Budgie1 said:

I'm not sure what software you use for processing but StarNet2 and StarXTerminator are both very good. Although they both allow you to remove the stars from the image, the stars are not lost, both of these also create an image with just the stars, which allows you to work on the stars and the background image separately. When you've finished processing both of them, you can put them back together again.

Working on background (nebula or galaxy) and the stars separately means you have more control over the adjustments you make. It also makes is easier if you want to use SHO or HOO for the main target and add RGB stars.  

I did not know this. I have used Star XTerminator for PS for a while and where is the image with just the stars? I use the original image for bringing back the stars (blend mode Lighten in PS).

I think it is a great image by the way Rodd! But have a go at trying starless processing. It really helps getting the weak signals out. I could not live without it now.

Edited May 30, 2022 by gorann

[Zummerzet_Leveller]

18 minutes ago, gorann said:

I did not know this. I have used Star XTerminator for PS for a while and where is the image with just the stars? I use the original image for bringing back the stars (blend mode Lighten in PS).

My understanding is that you use layers in imaging processing software and subtract your starless image from the image you created the starless from. This just leaves the stars.  You can then control star sizes, colour etc separately.  Pretty sure this isn't an automatic process, although it might be in pixinsight.

[Budgie1]

1 hour ago, gorann said:

I did not know this. I have used Star XTerminator for PS for a while and where is the image with just the stars? I use the original image for bringing back the stars (blend mode Lighten in PS).

I don't know about the PS version of StarXTerminator, but the PI version has a tick box marked "Generate Star Image" and in StarNet2 for PI it's called "Create starmask".

When used it creates a second image with only the stars and leaves the starless version in the original image.

When adding the stars back in, I just use PixelMath to add them together.

Edited May 30, 2022 by Budgie1

[Rodd]

1 hour ago, Budgie1 said:

I don't know about the PS version of StarXTerminator, but the PI version has a tick box marked "Generate Star Image" and in StarNet2 for PI it's called "Create starmask".

When used it creates a second image with only the stars and leaves the starless version in the original image.

When adding the stars back in, I just use PixelMath to add them together.

   The starless tool is no longer part of PI I don’t think. The script is no longer accessible.  I am using not the newest version, but almost 

[Rodd]

9 hours ago, vlaiv said:

Care to share the data ?

Here they are.  Several versions of the h stack.  46 has best FWHM, 90 has best signal.  The best fwhm is the o stack.

I just noticed star ringing in the nebula  darn it!  Back to the drawing board.  There is a decent image in there, I just need to find it

h46.fit

h80.fit

h90.fit

o65.fit

Edited May 31, 2022 by Rodd

[teoria_del_big_bang]

1 hour ago, Rodd said:

   The starless tool is no longer part of PI I don’t think. The script is no longer accessible.  I am using not the newest version, but almost 

Both StarXTerminator and Starnet (and Starnet2) are available still for PI.
Neither are scripts, as such, as they do not reside in the scripts menus, but modules. available under the Processes menu.

StarXTerminator sometimes does disappear when a new full release of PI is installed but you can easily get it back (assuming you have a licence as it is not free but a one off payment0.

• Occasionally PixInsight's update system gets confused. If anything goes wrong with the above procedure, or if StarXTerminator seems to install and work fine but then stops or disappears from the Process menu, please do the following:
  • Delete the "updates.xri" file from PixInsight's main application directory
  • Re-start PixInsight
  • This will force a fresh download/installation of all updates to your PixInsight installation, including StarXTerminator

StarXTerminator Support for PI

 

Starnet2 has now superseded Starnet.

Guide to install Starnet2

Steve

 

[Rodd]

2 minutes ago, teoria_del_big_bang said:

Starnet2 has now superseded Starnet.

Thanks Steve.  my computer and pi have always given me trouble.  I will look for these tools/processes as you suggest.  I hope I can find them!

[gorann]

Rodd, I had a go at your data using the h90 and o65 and started off as HOO and then tweaked the green towards blue as I am no fan of green in AP. After a series of initial stretches I made a starless version (with Star XTerminator in PS). That one was treated with the new RC-astro Noise XTerminator and then stretched and HiPass filtered (all in PS). I also used a bit of Topaz Denoise to sharpen it a bit and reduce the noise further. Then I brought the stars back at a reduced level (as a layer adjusted with a curve in PS). I also tried to fix the oblong stars in the corners a bit. Here are the starless and starred versions. I also added a less pink vesion.

Cheers, Göran

Edited May 31, 2022 by gorann

[Rodd]

38 minutes ago, gorann said:

am no fan of green in AP.

I have been wasting my time...I thought as much!!!  Wow.  But you forget that with certain nebula--supernova and PN, wolf ryat stars etc, teal is more accurate that pure blue.  Leaving the teal look was intentional.   Your image is definitally well processed.  I wish I could get data like that!!! (I guess I don't need more)--though I think a major tone down would be nice.  I do really like the background.  The nebula is a bit neon for me--though it sure looks nice from a detail and clarity prospective

The oblong stars were limited to the upper left hand corner and I do not know why they are there.  I rotated the camera and prior to rotation the stars were good across the frame.  

Pixinsight blows in my opinion--time and again I come to this realization.  But I simply refuse to pay a monthly charge for PS.  I have a disc with PS on it, but I really do not know where to start.  

IMO--if you put back maybe 15% of the green, and I think remove a bit of blue from the red portions of the nebula (it looks pink to me on my screen).  The background looks like the red I think is perfect..  But all in all--an amazing job of processing.  One thing I notice is there is more blue in teh background in the starless image--I rather like it--it has a translucent appearance like teh envelope around the nebula.  Its not present as much in the star image though.

I had not seen the less pink version--great minds think alike I guess.  Nice

Edited May 31, 2022 by Rodd

[Adreneline]

19 hours ago, Rodd said:

Very disappointed.  I do not understand it.

To be honest Rodd I don't really understand why you are disappointed. To my eyes this is an excellent image of a very difficult target. The main subject is well defined with a very pleasing amount of colour - not overstated but sufficient to reveal an extraordinary level of detail. As you pointed out to me on my recent M13 image backgrounds are tricky things and all the more so when the background itself contains nebulosity which again to my eyes looks controlled and well balanced withe star of the show - NGC6888.

As stated above I now use Starnet/StarXterminator as a matter of course in PI for processing nebula images; doing so provides a much greater degree of control over processing of the stars and the nebula. I know others say putting the stars back is a simple matter but that has not always been my experience if you want to avoid them looking 'stuck on'. Personally I never just put the processed 'star only' image back with PixelMath. Experimentation is best to find a method that works best for the image you are working on - in my opinion there is no single solution. I tend to use a duplicate which I process complete, taking lots of care not to overprocess and then use star reduction on that image and maybe MMT. I then remove the stars and blend those back using PixelMath and a 'lighten' formula or sometimes a 'screen' formula.

Adrian

[Rodd]

5 minutes ago, Adreneline said:

o be honest Rodd I don't really understand why you are disappointed.

Its hard to explain--this image (my version) is kind of right on the edge--at times I hate it becuase its noisy and cartooney, and other times I feel its not bad.  It certainly lacks the processing that Goran's has--its obvious my processing has fallen behind.   All the things you say about the image were very carefully striven for, so it pleases me to see someone has noticed.  As I said in an earlier post--there is a decent image in there and I just need to find it.  Goran found one.  

I will definitely see if I can find the starless tools.  I don';t really want to buy it though.  How much is it?  I will look for it in the process console.  I know it is not in my tools--you know you "list all" and then all the tools appear and you choose one--curves, historgram, TGV  etc.  Its not in that list.

What you say about puttying the stars back--sound about right.  Nothing in this venture is ever simply click and go.  The simplest task is best eased into after experimenting.   I hope I can find it. 

[Adreneline]

9 minutes ago, Rodd said:

I will definitely see if I can find the starless tools.  I don';t really want to buy it though.  How much is it?  I will look for it in the process console.  I know it is not in my tools--you know you "list all" and then all the tools appear and you choose one--curves, historgram, TGV  etc.  Its not in that list.

I purchased StarXterminator just before the new StarNet2 came out - wouldn't you know it!! Both are very effective but again both can give slightly different results - it all depends on the image.

Starnet and StarNet2 install as part of the PI installation - there is no additional cost.

(As an aside StarXterminator has now been joined by NoiseXterminator which seems to be generating a lot of interest on the PI front.)

13 minutes ago, Rodd said:

All the things you say about the image were very carefully striven for, so it pleases me to see someone has noticed.

There is so much subjectivity when looking at images - your own and images produced by others. We should all strive to get the best result from the data we have and in my opinion your image is up there with the best. Can it be improved? We should always like to think so  - otherwise what's the point?

Adrian

 

[Laurieast]

I have tried both StarNet2 and StarXterminator in PS and found them very similar. I found this YouTube video very helpful, and after watching it a couple of times, wrote down a work flow, which gets easier the more you use it.

https://www.youtube.com/watch?v=K08Qno7H0lw&t=5s

 

[Rodd]

1 hour ago, gorann said:

Rodd, I had a go at your data using the h90 and o65

And most impressive of all is its Bin 1--you did not bin teh data as I did.  Wow.  THAT takes skill, as it goes against the numbers (pixel sacle, seeing, guiding, resolution...all that jazz).  At full resolution your image is impressive.

[Rodd]

8 minutes ago, Laurieast said:

I have tried both StarNet2 and StarXterminator in PS and found them very similar. I found this YouTube video very helpful, and after watching it a couple of times, wrote down a work flow, which gets easier the more you use it.

https://www.youtube.com/watch?v=K08Qno7H0lw&t=5s

 

Thanks--I will give it ago.  i think my biggest challenge will be finding the tool and opening in the PI program.  Once I have the tool working, I will be able to figure it out pretty quickly........yeah, right!

[gorann]

Hi Rodd,

Adrian @Adreneline is right, putting the stars back is a bit tricky. They easily get dark ringed (panda eyes), which I think looks awful, but I found various ways to fix that. The easiest (in PS) is to select all the stars (using Select -> Color Range) and then expand the selection (usually about 3 pixels) and feather it (1 pixel). Then I use a curve on the selection to brightenn the area (ring) around the stars. It may be necessary to do it selectively on different parts of the image depending on the background (nebulosity of dark sky). There is probably a similar way to do it in PI.

Yes, Rodd, I was unsure about the colors. I did bring down the blue in the dark sky after I added the stars, but it can be easily adjusted up again if someone likes that better. I always found it difficult not to end up with a pink/mageta color on this nebula because of the Oiii shell around it. If I move it towards deeper red then the shell becomes less obvious also on the main part of the nebula and is mainly seen around it. Assuming it is the Oiii on top that Ha that turns the red to pink, maybe that is how it should be. Here is a tif file (16 bit since it is done in PS) that you can play with.

As you say the image would probably benifit from being tuned down a bit, making it less contrasty, I just wanted to see how far I could go.

Rodd NGC6888 HOO Gorans version.tif

[gorann]

Rodd,

maybe this one is more to your liking regarding the colors (tweaked the channel curves a bit).