M16 Pillars of Creation in mono

[Rodd]

On 6/14/2016 at 05:34, swag72 said:

This has got to be THE single most iconic image out there - So thanks to Hubble there's masses of pressure with this image! How can you improve on an icon? I have no idea at the moment! But the mono Ha data has come out OK!

This is a low one for me - Only gets to 37 degrees maximum and I was just starting to only image above 40 degrees  I will be collecting the OIII and probably SII as well (depends on how a bi colour comes out) - But meanwhile I thought that this stood on it's own two feet as a mono target. It was a delight to process as the data is nice and strong.

Details:
Mount: Mesu 200
Scope: Orion Optics ODK10
Camera: QSI683 with 3nm Ha Astrodon filters

28x1800s 14 hours in total.

You can see the larger res image here

Very smooth--which I assume you aficionados equate to low noise?  Still trying to identify that secret ingredient that yields noise free images without the need for noise reduction.  As everyone indicates, the image is spectacular, though to me it appears a bit pale.  I have seen some of your Ha images that are......less pale....brighter?  Trying to use the appropriate terminology but am falling short.  I don't want to use the term washed out, because that suggests a negative--and I am not intending to be negative.  I suppose if you don't know what I mean its not worth pursuing.  So, is it target specific, scope/camera specific, or a processing decision you made.  Or is it akin to me asking you why the ball looks so square? (which it doesn't)

[swag72]

@Rodd The data on this was great - It was as smooth as butter..... The secret? None from me, just 14 hours worth of data. I think I made a decision with this one to make it a little paler and 'fragile' looking.......... I could have upped the contrast, but I didn't want the Pillars to look dark and solid... if that makes sense. In all the images I've seen, it's like you can almost look through the pillars and I wanted to keep it like that

[ChrisLX200]

The Hubble makes the tallest pillar seem solid but in fact it's located deeper inside the nebula so you're viewing it through a haze of light from triple-ionized oxygen. It's always come out much softer and more delicate than the other two in the good images I've seen. It makes a big difference to the way it looks when the OIII channel is added :-)

ChrisH

[Rodd]

On 6/14/2016 at 05:34, swag72 said:

This has got to be THE single most iconic image out there - So thanks to Hubble there's masses of pressure with this image! How can you improve on an icon? I have no idea at the moment! But the mono Ha data has come out OK!

This is a low one for me - Only gets to 37 degrees maximum and I was just starting to only image above 40 degrees  I will be collecting the OIII and probably SII as well (depends on how a bi colour comes out) - But meanwhile I thought that this stood on it's own two feet as a mono target. It was a delight to process as the data is nice and strong.

Details:
Mount: Mesu 200
Scope: Orion Optics ODK10
Camera: QSI683 with 3nm Ha Astrodon filters

28x1800s 14 hours in total.

You can see the larger res image here

Great--then I'm not crazy.  Still convinced there is a missing element.  I just obtained auto focus capability--probably won't install it and master its use for some time--want to finish my the images I am working on first (Pacman, M27, and M97/M108).  Then I will see it that is the secret ingredient.  Not sure how focus impact noise though.

[Rodd]

6 minutes ago, ChrisLX200 said:

The Hubble makes the tallest pillar seem solid but in fact it's located deeper inside the nebula so you're viewing it through a haze of light from triple-ionized oxygen. It's always come out much softer and more delicate than the other two in the good images I've seen. It makes a big difference to the way it looks when the OIII channel is added :-)

ChrisH

I see the difference between the tallest one and the Eagle--as well as the other one on the left.  I hadn't noticed that.  Pretty cool.  

[Rodd]

1 hour ago, swag72 said:

@Rodd The data on this was great - It was as smooth as butter..... The secret? None from me, just 14 hours worth of data. I think I made a decision with this one to make it a little paler and 'fragile' looking.......... I could have upped the contrast, but I didn't want the Pillars to look dark and solid... if that makes sense. In all the images I've seen, it's like you can almost look through the pillars and I wanted to keep it like that

How do you keep the background ("black space between the stars") from appearing pixellated ("noisy"?).  

[swag72]

If there's no noise in the data it won't pixelate..... I did nothing other than stretching and some contrast enhancement and sharpening.

[Rodd]

1 minute ago, swag72 said:

If there's no noise in the data it won't pixelate..... I did nothing other than stretching and some contrast enhancement and sharpening.

Do you use liquid cooling? If not, then I am at a loss how to take a noise free exposure.  Perhaps that is the holy grail I have been after.

[swag72]

No I only cool to -15 degrees as I can attain that temperature all year round out here in Spain.

[Rodd]

2 minutes ago, swag72 said:

No I only cool to -15 degrees as I can attain that temperature all year round out here in Spain.

So my cooling to -40 or -45  is not the answer.  Hmm.

[ollypenrice]

39 minutes ago, Rodd said:

So my cooling to -40 or -45  is not the answer.  Hmm.

It won't do any harm but calibration, dither and hard work (one of Sara's more obvious strong points ) will give the same result.

Olly

[Rodd]

1 hour ago, ollypenrice said:

It won't do any harm but calibration, dither and hard work (one of Sara's more obvious strong points ) will give the same result.

Olly

Well I dither and calibrate, but from what Sara has indicated, her processing is light--no noise reduction, no masks, no pixel math, no blah, blah, blah.  I think she has a magic button labeled "perfection" 

[ollypenrice]

4 hours ago, Rodd said:

Well I dither and calibrate, but from what Sara has indicated, her processing is light--no noise reduction, no masks, no pixel math, no blah, blah, blah.  I think she has a magic button labeled "perfection" 

Sara takes plenty of data. That's her noise reduction - along with stretching carefully.

[Rodd]

15 minutes ago, ollypenrice said:

Sara takes plenty of data. That's her noise reduction - along with stretching carefully.

That is one of my points.  Even when I have 15 hours of data per channel, the noise is strikingly evident when compared to "good" images.  Quantity of data can't be the only thing.  

[swag72]

8 hours ago, Rodd said:

Well I dither and calibrate, but from what Sara has indicated, her processing is light--no noise reduction, no masks, no pixel math, no blah, blah, blah.  I think she has a magic button labeled "perfection" 

I will do a little noise reduction if needed - Generally perhaps just in the outer areas. I don't use star masks, pixel maths or any of that. Stuff is done in very small iterations as well, so if you look between the 23rd and 24th iteration for example there's almost no difference, but a few more iterations down the line and stuff will start to show. I keep changes along the way (once i get going) and I can have up to 50 saved images, so can go back to any of them and perhaps blend in a bit that got out of hand, or an area that I preferred.

[ollypenrice]

5 hours ago, Rodd said:

That is one of my points.  Even when I have 15 hours of data per channel, the noise is strikingly evident when compared to "good" images.  Quantity of data can't be the only thing.  

What's your sky like, Rodd? Do you have an SQM figure for it? I don't know about Sara's sky but I was on 21.71 last night. This makes a huge difference.

Sara's advice about lots of small iterations is right on the money. Adam Block makes the same point. I think it's widely accepted as the best way. Between each iteration be sure to look at the image at all scales - zoomed in and zoomed out - because some consequences are not visible in both.

Olly

[swag72]

11 minutes ago, ollypenrice said:

What's your sky like, Rodd? Do you have an SQM figure for it? I don't know about Sara's sky but I was on 21.71 last night. This makes a huge difference.

My skies aren't THAT good Olly - I measure perhaps 20.3 on a very good night........ If I'm looking south (as in M16) I am imaging into the skyglow of Spains 3rd city only 20km away  

[ollypenrice]

Just now, swag72 said:

My skies aren't THAT good Olly - I measure perhaps 20.3 on a very good night........ If I'm looking south (as in M16) I am imaging into the skyglow of Spains 3rd city only 20km away  

I knew you had LP, Sara, but I don't think I've ever heard you quote a figure before. I guess that in NB this has almost no significance? Do you DBE your NB data? I never do, but I sometimes do for Lum and always for RGB.

Olly

[swag72]

Is that a bad figure Olly? I have nothing really to measure it against and so just carry on merrily! Probably an indication of why I find LRGB imaging so hard and to be honest in the South I wouldn't even try it.

Never use DBE on narrowband data - in fact, don't use DBE at all now. Perhaps a little gradient exterminator on the LRGB but that's it.

[Rodd]

6 hours ago, swag72 said:

I will do a little noise reduction if needed - Generally perhaps just in the outer areas. I don't use star masks, pixel maths or any of that. Stuff is done in very small iterations as well, so if you look between the 23rd and 24th iteration for example there's almost no difference, but a few more iterations down the line and stuff will start to show. I keep changes along the way (once i get going) and I can have up to 50 saved images, so can go back to any of them and perhaps blend in a bit that got out of hand, or an area that I preferred.

 

4 hours ago, ollypenrice said:

What's your sky like, Rodd? Do you have an SQM figure for it? I don't know about Sara's sky but I was on 21.71 last night. This makes a huge difference.

Sara's advice about lots of small iterations is right on the money. Adam Block makes the same point. I think it's widely accepted as the best way. Between each iteration be sure to look at the image at all scales - zoomed in and zoomed out - because some consequences are not visible in both.

Olly

 

4 hours ago, swag72 said:

Is that a bad figure Olly? I have nothing really to measure it against and so just carry on merrily! Probably an indication of why I find LRGB imaging so hard and to be honest in the South I wouldn't even try it.

Never use DBE on narrowband data - in fact, don't use DBE at all now. Perhaps a little gradient exterminator on the LRGB but that's it.

Well--As far as my light pollution, I have no idea about SQM number--I will say I can see the milky way on a good, dark night, but it is faint.  M31 is never visible unaided.  Below 30 degrees does not exist.  I have been considering setting up in the Adirondacks (mountain park about 4 hours away) that has skies as dark as northern Canada (basically a 22 SQM).  But seeing as the Jet stream still passes over the region, I'm not sure if it is worth it.  I guess I don't know what is hurting me more, light pollution, or poor seeing and transparency.

Regarding iterations--this sounds magic buttony to me.  Thank you--I will pursue this approach.  PI has a tool that one can use to go back to any spot in the processing  history of an image.  Blending sounds like a PS thing.  I usually save an image after each process (despite having the history tool, which makes it unnecessary), and its not to long before I am swamped with images and confused.  I can't imagine how I will deal with saving 10-20 images PER PROCESS.  I have to come up with a better file saving system.  That's been a work in progress.

See--its is magic--no more DBE, masks, PM, noise reduction (for the most part) etc...etc.......pretty soon she won't have to do anything but snap the picture!

Processing aside--I went to a PI week long workshop during which we processed provided data.  The data was generated by other amateurs but looked so good it was almost unnecessary to do any processing.  Naturally I was able to process these images fairly well.  The starting point was very high.   I guess I have been trying to determine why my starting point is so low and perhaps light pollution is the main culprit? I will set up at a remote site for the final confirmation (something that I could not do until recently as Gremlins were the norm--now they are sporadic).

Thanks for the input

Rodd

[Rodd]

9 hours ago, ollypenrice said:

What's your sky like, Rodd? Do you have an SQM figure for it? I don't know about Sara's sky but I was on 21.71 last night. This makes a huge difference.

Sara's advice about lots of small iterations is right on the money. Adam Block makes the same point. I think it's widely accepted as the best way. Between each iteration be sure to look at the image at all scales - zoomed in and zoomed out - because some consequences are not visible in both.

Olly

Here's a question for you Olly--I know that stacking 30 subs reduces noise compared to stacking 5 subs--the more subs the higher the SNR to a point.  Does this hold true for subs of different filters?  If I stack 5 subs of each LRGB to have 20 subs, will the SNR be equal to 20 subs of, say., the Lum channel?  Or is SNR calculated per filter and then added, so a channel with 5 subs will add allot of noise to a final image even if that image has 30 subs each of all other filters?

[steppenwolf]

I also like iterative processing with very small changes each time although I tend to break mine down into sets of multiple layers until I reach about 6 layers and if I am happy thus far, I save that set then flatten the image and use that as the base for the next 6 layer stack. This gives me ample opportunity to go back if I need to or blend in data from an earlier rendition. Processing cannot be rushed as Sara will no doubt testify!

[Rodd]

7 minutes ago, steppenwolf said:

I also like iterative processing with very small changes each time although I tend to break mine down into sets of multiple layers until I reach about 6 layers and if I am happy thus far, I save that set then flatten the image and use that as the base for the next 6 layer stack. This gives me ample opportunity to go back if I need to or blend in data from an earlier rendition. Processing cannot be rushed as Sara will no doubt testify!

I assume from your language you use PS?

[Rodd]

16 minutes ago, steppenwolf said:

I also like iterative processing with very small changes each time although I tend to break mine down into sets of multiple layers until I reach about 6 layers and if I am happy thus far, I save that set then flatten the image and use that as the base for the next 6 layer stack. This gives me ample opportunity to go back if I need to or blend in data from an earlier rendition. Processing cannot be rushed as Sara will no doubt testify!

Just ordered your book--Thanks for the link.  I have been meaning to do it for some time.  

Rodd

[steppenwolf]

Quote

I assume from your language you use PS?

Yes, indeed I do.

Quote

Just ordered your book--Thanks for the link.

Thank you, I hope you enjoy it!