M17 Data Set

[gorann]

8 hours ago, wimvb said:

I tried with a different palette as well. But for some reason I just can't get the same amount of detail out of it using a more "naturally" looking palette.

Anyhow, here's a SHOO version of the data (actually S.7H.3OO.9H.1)

Click for a larger version (not that it looks any better)

I still like it alot. I assume that using a Hubble palette has the advantage of clearly separating the info in the different filters and therefore more structural information is shown, Lumping the red wavelengths Ha and Sii into the red channel must mean that visual information is lost, but I like the more natural look. I have never been a fan of all that green in NB images - there is almost no green in the real universe (unless you zoom in on planets like the Earth ?? ). One possibility that now struck me is to make a traditional Hubble image and then make it monochrome and use it as lum on an image like your last one. Maybe you could try it since you have done both versions.

[gorann]

Wim, since I am rather new to the jargon of NB processing, can you explain what "S.7H.3OO.9H.1" means?

[wimvb]

Just now, gorann said:

Wim, since I am rather new to the jargon of NB processing, can you explain what "S.7H.3OO.9H.1" means?

Red channel: 0.7 x Sii + 0.3 x Ha

Green channel: Oiii

Blue channel: 0.9 x Oiii + 0.1 x Ha

As you wrote earlier, mixing probably ruins the structural data. I'll make a luminance out of a mix if Ha, Oiii and Sii, and apply that to the latest version.

[Rodd]

12 hours ago, wimvb said:

Anyhow, here's a SHOO version of the data (actually S.7H.3OO.9H.1)

That's quite a palette descriptor.  I like it allot.  

Ridd

[gorann]

So I just had to try my idea of using a Hubble palette (Ha to green, Sii to red and Oiii to blue) as a lum layer in an effort to add more detail to my RGB like palette. In addition I had noticed that some of the images produced by the PI guys in this thread showed more details than mine, so I attacked the Hubble palette image with some deconvolution and HDRM in PI. I think it worked, since there is more detail in the image now.

So to summarize, this is a version where I first used Skarpen's palette (with most of the Ha and Sii in the red channel) and then used a Hubble palette version as a lum layer in PS (after first making it monochrome).

Then, of course, I could not stop myself from playing with curves and unfortunately came up with two versions I could not decide upon, so there they are. In the second one there is more amphasis on the dark nebulosity but it may be over the top.

 

[Rodd]

31 minutes ago, gorann said:

Then, of course, I could not stop myself from playing with curves and unfortunately came up with two versions I could not decide upon, so there they are. In the second one there is more amphasis on the dark nebulosity but it may be over the top.

Now you are turning into me...the difference is so slight, others probably have a hard time really seeing it without undo scrutiny.  They are both excellent!

Rodd

[gorann]

1 minute ago, Rodd said:

Now you are turning into me...the difference is so slight, others probably have a hard time really seeing it without undo scrutiny.  They are both excellent!

Rodd

Yes, I can see it is difficult to see. Sorry to confuse you. It was more apparent in PS when I was flipping between the versions. In any case I go for the last one and has just put it on Astrobin.

[Rodd]

Just now, gorann said:

Yes, I can see it is difficult to see. Sorry to confuse you. It was more apparent in PS when I was flipping between the versions. In any case I go for the last one and has just put it on Astrobin.

Well done, and no worries

Rodd

[wimvb]

I agree, @gorann. Nr 2 has slightly more in the darker parts.

Anyway, here are two versions from me. Used the image from yesterday and added luminance created from combining the Ha, O and S masters. Then processed this luminance and used LRGB-combination. One version with HDR held back, and one "classic" PixInsight. Not as subtle a difference as in Göran's version. Which do you prefer?

 

 

 

[gorann]

Great images Wim and amazing structure you brought out from that blue centre! I must have blown some of it out during my stretch. I should have a new go at it tomorrow. Great learning excercise

[Rodd]

2 hours ago, wimvb said:

I agree, @gorann. Nr 2 has slightly more in the darker parts.

Anyway, here are two versions from me. Used the image from yesterday and added luminance created from combining the Ha, O and S masters. Then processed this luminance and used LRGB-combination. One version with HDR held back, and one "classic" PixInsight. Not as subtle a difference as in Göran's version. Which do you prefer?

 

 

 

I like the first one I think. Well done

Rodd

[wimvb]

46 minutes ago, Rodd said:

I like the first one I think. Well done

Rodd

Me too. But the other version does have some nice areas surrounding the blue region. I could use a colourmask and blend the two. Or not; decisions, decisions.

[MilwaukeeLion]

Thanks for awesome test data Rodd.  Only captured and attempted my own narrowband data combination with Dumbbell which was less difficult than this. 

My attempt processing your narrow band data in SHO - I can either get the nebula close to right color but stars are purple or get stars close to right color but nebula is completely green using Nebulocity LRGB color synthesis.  So I did one of each and merged them best I could in Photochop.  This is why im still hesitant buying cooled ccd, skills with data...

[Rodd]

5 minutes ago, MilwaukeeLion said:

Thanks for awesome test data Rodd.  Only captured and attempted my own narrowband data combination with Dumbbell which was less difficult than this. 

My attempt processing your narrow band data in SHO - I can either get the nebula close to right color but stars are purple or get stars close to right color but nebula is completely green using Nebulocity LRGB color synthesis.  So I did one of each and merged them best I could in Photochop.  This is why im still hesitant buying cooled ccd, skills with data...

That's nice looking.  You will get the hand of a cooled CCD (or CMOS).  It does take a bit of work.  The learning curve can be frustrating.  The reward or course, is watching this amazing window to the universe clear up and come to life before your eyses.

Rodd

[gorann]

Before moving onto Rodd's new challenges I have had a final go at M17. I think I just about managed to get it as detailed as Wim by some HiPass filtering and careful curve adjustments. However, it struck me that this version then became a bit "in your face", maybe too sharp and contrasty. This was saved version 37 and I then had a look at my version 2 (about an hour of processing compared to two days) and was slightly chocked how good and soft it looked. It made me wonder if I just should have stopped there and not chased to much detail. Well, I like both sofness and detail so I added 40% of version 2 to version 37. So here they are, the image I sharpened as much as I could, the old version 2, and finally a mix of them.

Any thoughts?

[Rodd]

17 hours ago, wimvb said:

Me too. But the other version does have some nice areas surrounding the blue region. I could use a colourmask and blend the two. Or not; decisions, decisions.

Tghe other one has more faint nebulosity around teh main event--including the blue emissions to the left.  But teh stars are not as clear--they are colored, which threw me a bit at first.  If I look to long I start to waffle between teh images.  Both are good.

Rodd

[Rodd]

1 hour ago, gorann said:

So here they are, the image I sharpened as much as I could, the old version 2, and finally a mix of them.

Any thoughts?

I do that allot too--make mixtures.  I agree that the first one is too angular and sharp--ungassy looking.  The combo is nice.  Its weird when one's processing that the line of "going to far" is so easy to cross without one knowing it.  It looks like a line from afar, but when you get to it, its a mile across.  You take a look behind you and you notice that you have suddenly left the farm........a long time ago!  Stepping away definitely helps.

Rodd

[gorann]

Yes Rodd, this is not the first time it happened to me and I nowadays store every litte step so that I can go back when I have drifted off. You talk about blue nebulosity to the left. I may have found some of it by tweaking the curve of the blue channel. It looks more like blue dust than nebulosity but it may look rather good.

[Rodd]

1 hour ago, gorann said:

It looks more like blue dust than nebulosity but it may look rather good.

I might be wrong, but cold dust typically is dark, while blue, reflection nebula is gas excited by reflected light.  So I think it is more nebulosity than dust.  Probably both.  But--I could be in error.

Rodd

[wimvb]

3 hours ago, Rodd said:

I do that allot too--make mixtures.

 

2 hours ago, gorann said:

I nowadays store every litte step so that I can go back when I have drifted off

That's the beauty of working only in PixInsight; I save my process as a project, so I don't have to save intermediate images. But otoh, a process can be up to 10 GB. The real beauty is when I want to redo a step. I just back up to where I want to insert a step, create a clone, and apply that step to the clone. Then I apply the rest of the process of the original image to the clone. Reworking an image can take only a few minutes. This doesn't work if you switch between software of course.

2 hours ago, gorann said:

It looks more like blue dust than nebulosity but it may look rather good.

 

1 hour ago, Rodd said:

I might be wrong, but cold dust typically is dark, while blue, reflection nebula is gas excited by reflected light.  So I think it is more nebulosity than dust.

In my image, the area is strong in blue, which corresponds to oxygen. Otoh, a filter doesn't differentiate between atoms, it differentiates between wavelengths. So the signal may very well come from a broadband source such as a reflection nebula. The blue light is then the light reflected from young, blue stars, rather than emitted from oxygen ions.

The various renderings in this thread made me wonder where the stars are that create the light (even the light from emission nebulae must ultimately come from some strong stars that excite the atoms, so they can in their turn emit light). Normally an emission nebula such as Ha, receives its light from a young star cluster within it. In this region, I don't see such a cluster. Or is it hidden within the pear shape that is M17? it almost looks like light is coming from within. What do you think?

[Rodd]

4 minutes ago, wimvb said:

The various renderings in this thread made me wonder where the stars are that create the light (even the light from emission nebulae must ultimately come from some strong stars that excite the atoms, so they can in their turn emit light). Normally an emission nebula such as Ha, receives its light from a young star cluster within it. In this region, I don't see such a cluster. Or is it hidden within the pear shape that is M17? it almost looks like light is coming from within. What do you think?

Distances can be deceiving as well--a star we think of as close or far can be the opposite, as size (brightness) varies.  Is there a cluster or hot star just off the frame?  I don't know.  There certainly is enough density in the nebula, even the non blue regions, to hide an emitting star.  Maybe the stars inside the blue area are heating up the blue region so fiercely that the blue internal nebula is emitting strongly enough to light the outer blue regions.  Or, maybe there is drift--the blue gas we see to the left may have originated in the center but was blown outward by solar wind to its current position.  Once heated, gas can remain excited for long periods--like portions of the Veil.

Rodd

[gorann]

I assmue you have read that Wiki says that "The open cluster NGC 6618 lies embedded in the nebulosity and causes the gases of the nebula to shine due to radiation from these hot, young stars; however, the actual number of stars in the nebula is much higher - up to 800, 100 of spectral type earlier than B9, and 9 of spectral type O, plus over a thousand stars in formation on its outer regions. It is also one of the youngest clusters known, with an age of just 1 million years." But as you say, the whereabout of these stars are not obvious.

[wimvb]

7 minutes ago, gorann said:

I assmue you have read that Wiki says

No, actually. But glad that you did. One less mystery to solve.

[wimvb]

I was never really happy with the SHO version of this image, so I reworked it. I don't know which version this is, so let's call it v 3 1/3

[Rodd]

2 hours ago, wimvb said:

was never really happy with the SHO version of this image, so I reworked it. I don't know which version this is, so let's call it v 3 1/3

I prefer the first image of your double post a while back.  This one seems wan, and for me a wee green heavy.  Its nice of course--stars are great, as is the background.  The detail is all there, it just seems a bit flat in the contrast department.  One tweak could totally change things.

Rodd