NGC 6888 HOO - a dissapointment

[Rodd]

5 minutes ago, gorann said:

I just wanted to see how far I could go.

You've gone at least as far as the Cresecnt Nebula--and that is a heck of a long way (5,000 light years).  Unfortunately, when you get back, depending on how fast you travel, we all will either be dust and there will be no one left to amaze, or we will not have realized you left at all.  We will ask you "did you forget something?" when you step out of your space craft upon returning.  You will look at us like we are crazy becuase you had been gone for a long time and had seen many womnderful things.  But to us--you hadn't left yet.  There was a great SF story about this  "Stardust"?  Can't remember the name--1950s era--the golden age.  In fact I think it was in an anthology by Asimov called "Before the Golden Age".  3 scientists invent faster than light travel and two go out exploring.  Upon their return, they were watching in a telescope and the pilot sees his wife having an afair with Gunderson--the other scientist.  Time was moving backwards for them, and when they touched down, they were right back at the point where they left.  So when the pilot jumped out of the ship all were wondering why--for they were just about to leave.  Then the scientist blugeoned Gunderson and murdered him, all the while ranting about a strange society they found and the afair that had yet to happen.  He was deemed crazy and locked away.  No one could explain the faintly sparkling dust that was on the nose of the brand new space ship that had yet to leave Earth.

This is what happenes when one processes data for too long without a break....they mumble on about completely irrelevant things!

[Rodd]

6 minutes ago, gorann said:

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

Now we are set upon a fools erand--trying to please my perfectionist tendancies.  You are, no doubt, aware of my penchant for having 30-40 versions on my Astrobin pages?  The same will happen here, if we persist.  Your image is very nice--representing far more processing skill than mine.   Many of the things I would change are a metter of personal opinion.  Others can't be changed without ecffecting other aspects that one does not wish alter--so what follows is a flurry of steps tryimng to undo that impact.   Its your image, and suited for your collection.   I guess all that lack of darkness has really fueled your processing ability.  stunning.

[gorann]

1 minute ago, Rodd said:

Now we are set upon a fools erand--trying to please my perfectionist tendancies.  You are, no doubt, aware of my penchant for having 30-40 versions on my Astrobin pages?  The same will happen here, if we persist.  Your image is very nice--representing far more processing skill than mine.   Many of the things I would change are a metter of personal opinion.  Others can't be changed without ecffecting other aspects that one does not wish alter--so what follows is a flurry of steps tryimng to undo that impact.   Its your image, and suited for your collection.   I guess all that lack of darkness has really fueled your processing ability.  stunning.

With astrodarkness gone up here now I could of course not resist your data - I knew it would be good as many times before, so thanks for posting it Rodd!

[Rodd]

12 minutes ago, gorann said:

With astrodarkness gone up here now I could of course not resist your data - I knew it would be good as many times before, so thanks for posting it Rodd!

My pleasure--You have shown me there is hope, after all, and that more data won't necesarilly be enough to cross the finish line.

[vlaiv]

Late to the party with my process, but I do hope you like it. I went for natural looking stars, airy looking nebulosity and as much of detail as I could pull out

ImageJ / Starnet v2 (standalone) / Gimp

[Rodd]

21 minutes ago, vlaiv said:

but I do hope you like it

Indeed I do--very natural looking.  It resembles a better version of mine.  Love the background.  I think both images (yours and Gorans) have shown me that there is ample data.  More, naturally, would be good, but only if it was as good in quality as this data--and with my sky, that is never a sure thing.  But even if I do manage to collect more and more good data--creating an image of quality is still not a forgone conclusion.  Bottom line.....I have to put my nose to teh grindstone and improve my processing skills!     

And you managed seperation in this pair of stars--Wow--that is the first I have seen.

                                                                                                                                                                                                          

Edited May 31, 2022 by Rodd

[gorann]

4 hours ago, vlaiv said:

Late to the party with my process, but I do hope you like it. I went for natural looking stars, airy looking nebulosity and as much of detail as I could pull out

ImageJ / Starnet v2 (standalone) / Gimp

I like your version a lot Valiv! Detailed, soft and spoky! Why did you downsample it before posting? I like to zoom in, especially here where we try to see what we can get out of the data and learn from that.

Edited May 31, 2022 by gorann

[vlaiv]

Just now, gorann said:

I like your version a lot Valiv! Detailed, soft and spoky! Why did you downsample it before posting? I like to zoom in.

Well, in order to get good tight stars and good detail I needed SNR and I don't particularly like over sampled look when fully zooming in, so I binned data x3.

Binned x3 gives about 1.215"/px (quick measurement) which is just about right for this data to make it look sharp at 100%. Additional SNR makes it a bit easier to sharpen properly without introducing too much noise.

 

[Ags]

@vlaiv I really like your process - a refreshing take on a familiar target.

[Rodd]

9 minutes ago, gorann said:

I like your version a lot Valiv! Detailed, soft and spoky! Why did you downsample it before posting? I like to zoom in, especially here where we try to see what we can get out of the data and learn from that.

Normally when shooting at .4 arcsec/pix and one achives a mean FWHM of 2.2-2.5, binning is called for as no resolution will be lost and the signal will be boosted.  But when you have processpors ab;le to do what Claiv and you can do--perhaps binning is not needed.  Bothwould hld up well to Bin 1--yours we see, Vlaivs we have to imagine--but its not hard to do

[Rodd]

6 minutes ago, vlaiv said:

Well, in order to get good tight stars and good detail I needed SNR and I don't particularly like over sampled look when fully zooming in

A couple of questions--how does binning tighten up satrs?  And, what is an oversampled look?  I am not sure what you mean.

[vlaiv]

Just now, Rodd said:

Vlaivs we have to imagine--but its not hard to do

Don't really know if I could pull out that much sharpness if I did not bin.

I'm not saying it can't be done - but it's certainly above my skill level. Enough SNR just makes it possible for me.

[Rodd]

4 minutes ago, vlaiv said:

Don't really know if I could pull out that much sharpness if I did not bin.

I'm not saying it can't be done - but it's certainly above my skill level. Enough SNR just makes it possible for me.

I have never heard it put this way--that binning can actually increase sharpness.  I guess if sharpening is facilitated by SNR, then one can sharpen bin3 more than Bin1---since there is no data loss (no resolution loss) after binning--I guess it makes sense.

[gorann]

2 minutes ago, vlaiv said:

Don't really know if I could pull out that much sharpness if I did not bin.

I'm not saying it can't be done - but it's certainly above my skill level. Enough SNR just makes it possible for me.

I do the opposit approach - I down sample when I am done with the image to a scale that does not reduce resolution. I tried a few times working on downsampled raw images and was not convinced it helped in any way, but....

[Rodd]

5 minutes ago, gorann said:

I do the opposit approach - I down sample when I am done with the image to a scale that does not reduce resolution. I tried a few times working on downsampled raw images and was not convinced it helped in any way, but....

The problem with downsampling after you have stretched teh data is you have lost liniarity.  The binning (downsampling) process works best wjhen teh data is linear.  I think you are using it just to get a size reduction.  But teh real benefit of binning is increasing SNR

[vlaiv]

4 minutes ago, Rodd said:

A couple of questions--how does binning tighten up satrs?  And, what is an oversampled look?  I am not sure what you mean.

When you bin you increase SNR - that part is easy.

In order to split stars from nebulosity, Starnet expect 16bit data - which I really don't like - I like to process images at 32bit per channel, so in order for it to do its thing, I need to "pre-stretch" image.

With binned data - more faint stars are kept as they show up in initial pre stretched image and can be distinguished from the noise.

That is first part.

Second part is in how much you can sharpen your data. Although I do selective sharpening - I do like high SNR as much less noise is generated / brought up to the surface by sharpening process and it is easier to control. This makes possible for tighter stars and more detail in nebulosity as well.

Oversampled look - well, compare the look of these three stars - just central core:

This is same star at three different bin levels (or sampling rates). Each of these is 100% zoom.

I like when stars stay points even at 100% and don't turn into little (or big) balls.

 

[Budgie1]

Even later to the party but I didn't get time to have a go until this evening.

Anyway, here's what I managed to do with the data using h46 & o65 and put through PI. The stars are not very good, the colour is okay but they're a bit eggy.

I used StarXTerminator to separate the stars so I could work on the nebula and spent more time on that than the stars, so maybe should have done more with those.

[Rodd]

Wll--here is my revision.  I tried to figure out as much as I could from Vlaiv's version--the extremely tight stars, the sense of reality--a depth that is hard to explain.  Not in your face, and teh detail.  But I a mystifyed.  I ended up about where I started--maybe a bit better.  I truied starnet in PI, but I do not know what RGB weight files are--they are required to use the tool and I dot know how to create them.  there is another file needed as well but I can't recall what it is.  Maybe if I watch teh tutorial.  I did bin 3, using teh 1.23' as the best reolution for the data.   I am not 100% sold on that--my first version was Bin 2.  I do like a bit of a zoom element.  But teh increased signal to noise is nice too.  The definition of crazy is repeating the same steps and expecting different results.  The problem is my tool box is empty....I have no choice.

I added a more saturated version, and a less bright veraion (with more color....AHHHHHHH!  Not sure which I like better

 

Edited June 1, 2022 by Rodd

[Rodd]

On 30/05/2022 at 19:48, 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".

Martin--I found Starnet on PI--but in order to use it I have to input RGB profiles (reference images of some kind) and there is a secnd reference image I have to upload--they are sprecific images--like drixzzle files.  You click on the spce and the software takes you to the folder where they are stored.  I do not know what these are or how to create them.

Rodd

[Budgie1]

1 hour ago, Rodd said:

Martin--I found Starnet on PI--but in order to use it I have to input RGB profiles (reference images of some kind) and there is a secnd reference image I have to upload--they are sprecific images--like drixzzle files.  You click on the spce and the software takes you to the folder where they are stored.  I do not know what these are or how to create them.

Rodd

The files should be in the PixInsight program folder, here's where mine are and what they're called:

The original StarNet has now been replaced with StarNet2, which I linked too on the first page of this thread. This is a much improved version on the original, so I would download that one. I think Steve - @teoria_del_big_bang also put a link to a tutorial for how to install StarNet2 in PI.  

Edited June 1, 2022 by Budgie1
Spelling

[Rodd]

1 hour ago, Budgie1 said:

The files should be in the PixInsight program folder, here's where mine are and what they're called:

The original StarNet has now been replaced with StarNet2, which I linked too on the first page of this thread. This is a much improved version on the original, so I would download that one. I think Steve - @teoria_del_big_bang also put a link to a tutorial for how to install StarNet2 in PI.  

Thanks Martin

[tomato]

Apologies Rodd if it was never your intention to allow a free for all on processing your data, but here is an offering processed in Startools. It is obviously far away from the previous images' colour palette but I have posted it as a testament to the quality of your data. I much prefer this to any version I produced with the IKI 80 hrs of SHO data. I can see structure reminiscent of the Spaghetti Nebula which I had never noticed in NGC 6888 before.

Thanks for posting the data. 

 

[Rodd]

41 minutes ago, tomato said:

Apologies Rodd if it was never your intention to allow a free for all on processing your data

That was the intention, the more the merrier.  I am glad you found it worthy.  You image is quite detailed.  Did you add SII from another data set?  It looks like a SHO palette, with the Ha regions gold.  Or did you use only my Ha and OIII and varying percentages and additions?   I tried to make a superluminance out of the Ha and OIII, but I found adding any data as L worked to deteriorate the other channel.  When I added Ha, the OIII became very dim.  When I added OIII, teh Ha receded.  I usually make HaSHO images when I do Hubble Palette images--invariably I add Ha as luminance and it always improves the image.....not so much with bicolor images though.  I found teh same true for the veil.

[tomato]

I used only your Ha (H46.fit) and OIII data, assigning the Ha to the green and the OIII to the blue in the Compose module of Startools. There is a vast choice of pre-set channel matrices in the software, I chose the 'Interpolate G 100R, (50R+50B), 100B' option.

The image was a little noisy after the sampling SVDeconvolution exercise in Startools, so I applied the NoiseXTerminator module in PI, set at 50% strength and 50% detail as a final operation.

[Rodd]

22 minutes ago, tomato said:

NoiseXTerminator module in PI, set at 50% strength and 50% detail as a final operation.

I don't have noiseXterminator.  Is that something one has to add to PI?