Jump to content


Photo

My M63

  • Please log in to reply
26 replies to this topic

#1 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 06 May 2013 - 11:11 AM

Conditions were nearly perfect for me, and I thought the pictures themselves turned out pretty well, my editing on the other hand almost makes me want to give up this hobby. I've tried for days in PI and just can't seem to get a handle on it. I don't exactly know why either, I've done other work in PI that I thought looked pretty descent, but my pictures from this last Friday and Saturday night look like complete garbage. Not happy with this in the least bit, but I was hoping by posting someone might have a clue as to where I'm going wrong.

I guess one of my main issues is how to properly darken the background of an image, I just can't seem to figure it out, well, I did, but clearly not the correct way IMO. Also, my stars look like *BLEEP*, although they were nice and pinpoint in my individual frames.

Posted Image

#2 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 06 May 2013 - 11:31 AM

Here is a dropbox download link to my DSS autosave, straight out of DSS with no modification or calibration:

https://dl.dropboxus...m63Autosave.tif

#3 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 12:06 PM

I'll have a look now

#4 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 12:32 PM

I cannot download via dropbox...dunno why.
Can you use another file sharing?

#5 nwinston

nwinston

    Ranger 4

  • *****
  • Posts: 333
  • Joined: 10 Feb 2010
  • Loc: Maryland, USA

Posted 06 May 2013 - 05:44 PM

I think that your problem with the stars lies in the processing. Like you said, I think that the strange star problem occurred when you tried to darken the background.

I gave the image a try at processing, and your data looks good. Here's my attempt at processing your image with Photoshop CS5.

Here's a link to the full size JPEG.

Attached Files



#6 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 05:45 PM

How did you download it from dropbox?

#7 nwinston

nwinston

    Ranger 4

  • *****
  • Posts: 333
  • Joined: 10 Feb 2010
  • Loc: Maryland, USA

Posted 06 May 2013 - 05:47 PM

I clicked the dropbox link and the file began to download automatically. What browser are you using?

#8 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 06:22 PM

Chrome: it appears the quick time icon in the centre of the page and nothing happens :). Same thing with Mozilla...
I'll try to disinstall QT...

#9 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 06:45 PM

Ok now it is downloading. Just had to remove QT :(

#10 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1922
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 06 May 2013 - 09:06 PM

Darkening the sky background is a personal preference. Some folks like it dark others like it lighter. You shouldn't clip it though.

Are you using a starmask to protect your stars as you stretch the histogram or use other tools like HDR compression, or contrast enhancement? Until I learned to do this my stars were messed up too - halos sometimes and doughnuts at other times.

Ajay

#11 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 06 May 2013 - 09:27 PM

I made a very quick elaboration.
It is on my working area in astrobin.
Here: http://www.astrobin.com/41159/

Your elaboration process must be reviewed, I think.

Regards,
Roberto

#12 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 07 May 2013 - 12:38 AM

I am using a starmask when doing the HDR tools and Contrast adjustment, however my *BLEEP* looking stars, I think, occured from AtrousWaveletTransformation. Perhaps I will re-do my processing leaving this out and just doing NR via ACDNR and see what happens.

Those colors are amazing Rob, I dunno how you managed to pull those out, as mine was at full +10 saturation boost and I could not even see those vivid of colors. Perhaps I had bad color calibration? I'm not really sure how to do color calibration via any other means then the PI "Galaxy Method" which I would guess most know what that process is. Perhaps I don't need to do color calibration beyond using the DBE tool which, as I understand, does a pretty descent calibration on its own? Just not sure.

Also Rob, you don't need to uninstall QT. Just right click the link, save target as, and save it. Should work that way I believe.

#13 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 07 May 2013 - 06:52 AM

Hi Sean,
the steps that I basically used are the following [in PixInsight]:
[-] Linearity [-]
1.
Cropped since your flat files were not so good.
Have a look at here: http://postimg.org/image/4j79zw0tv/
2.
DBE for reducing some gradients
3.
Background Neutralization
4.
Color Calibration

[-] Non Linearity [-]
5.
Creation of the mask on the galaxy and mask on the stars
6.
Pixel Math to create a combo mask --> mask galaxy | mask stars
7.
7.1 Apply the combo mask with galaxy and stars blocked.
7.2 SCNR on the green channel (strength: 1.0)
8.
8.1 Apply the combo mask with the background blocked.
8.2 Extract the lightness of the image
8.3 LRGB combination [de-checking R-G-B and select for L the lightness of point 7.2] with saturation around 0.2
9.
Local Histogram Equalization

I think that's all.
Regards,
Roberto

#14 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 07 May 2013 - 07:02 AM

I give you a - strictly personal - hint: try to use as less as possible the noise reduction (ACDNR or ATrous or whatever).
I do not like at all the effect it comes out; I mean, not in PixInsight specifically, but in all the tools.
I find that noise reduction gives to the image an innaturale appearance.
It is preferrable, for my POV, to increase the integration frame set.

Regards,
Roberto

#15 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 07 May 2013 - 08:36 AM

Thanks for the hints. I agree about NR, I'm not a big fan of it either.

My flats I thought were just about as good as you can get, the problem however is my scope is a 6" SCT and I'm shooting with a large DSLR sensor, so the vignetting can't really be helped, I just have to crop, or so I thought. Is this really the case, or am I truly shooting bad flats? From the looks of it my flats look pretty similar to examples posted by other people.

#16 rob77

rob77

    Viking 1

  • -----
  • Posts: 566
  • Joined: 25 Feb 2012
  • Loc: Belo Horizonte (Brazil) and Cuneo (Italy)

Posted 07 May 2013 - 08:39 AM

Could you post a RAW flat here?

#17 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 07 May 2013 - 10:33 AM

Here is the dropbox link to a RAW flat from the M63 photo:

http://db.tt/YY71PUqz

I just use the laptop screen in front of the scope method and the AV/Flat setting in Backyard EOS to do my flats. I thought they looked pretty normal considering the small scope and large camera sensor. Let me know what you think. Thanks!

#18 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 07 May 2013 - 11:33 AM

I now see that the reason I was over-darkening my background was because I didn't use the SCNR tool to remove the color, which led me to over-darken to get rid of it instead.

#19 Nils_Lars

Nils_Lars

    Voyager 1

  • *****
  • Posts: 10055
  • Joined: 04 Jan 2008
  • Loc: Santa Cruz Mountains , CA

Posted 07 May 2013 - 12:13 PM

A good shot of the Sunflower, processing is always a trick.

#20 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1922
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 08 May 2013 - 12:25 AM

I am using a starmask when doing the HDR tools and Contrast adjustment, however my *BLEEP* looking stars, I think, occured from AtrousWaveletTransformation. Perhaps I will re-do my processing leaving this out and just doing NR via ACDNR and see what happens.


I prefer Atrous Wavelet Transformation because you hit he noise before you stretch. A few things that have worked for me:
* Apply noise correction to just two layers - i.e. at the two smallest scales.
* Since you are correcting the background, protect the high signal region with a lightness mask.
* Another trick is to reduce star size via Morphological Transform before stretching. This keeps down star bloat in later steps particularly after stretching.

Have fun.

Ajay

#21 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 08 May 2013 - 01:42 PM

Thanks for the info. I was using Atrous on all layers, I will reduce it. Also, I've used Morphological Transformation before but always in non-linear form, I will try using it before the stretch.

I think my major issue is that I'm bad as masking. Those darn things take so much tweeking to be perfect. I'm having a bit of trouble with the star mask too, by the time I get it to include all stars its picking up a lot of extra noise twords the edges of the image as well. Its so hard to do without being able to see a real time preview.

#22 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1922
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 08 May 2013 - 11:10 PM

A quick way to make a luminance mask is as follows:

* make a clone and stretch it quickly using STF
* open the ACDNR tool and select the preview of the lightness mask and tweak it. Of course you will be seeing an inverted image
* When satisfied **with the preview still selected** drag and drop the instance on the **stretched clone**.
* this will give you an mask you can use on your image

Cheers,
Ajay

#23 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 08 May 2013 - 11:24 PM

In my lum mask, do I want the background to be entirely white? Should I be tweaking it until only stars and structures are visible with a completely white background?

#24 bluedandelion

bluedandelion

    Surveyor 1

  • *****
  • Posts: 1922
  • Joined: 17 Aug 2007
  • Loc: Hazy Hollow, Western WA

Posted 09 May 2013 - 12:11 AM

Correct. You want the background to be white if you are using an inverted mask. As to how much structure you want visible, depends on the process you are doing, but for background noise reduction you should make all stars and structures visible.

Ajay

#25 Sean13

Sean13

    Mariner 2

  • -----
  • Posts: 205
  • Joined: 17 Nov 2012
  • Loc: North Platte, Nebraska

Posted 09 May 2013 - 10:29 PM

I prefer Atrous Wavelet Transformation because you hit he noise before you stretch. A few things that have worked for me:
* Apply noise correction to just two layers - i.e. at the two smallest scales.
* Since you are correcting the background, protect the high signal region with a lightness mask.
* Another trick is to reduce star size via Morphological Transform before stretching. This keeps down star bloat in later steps particularly after stretching.

Have fun.

Ajay


Sorry a couple more questions here. When you say only on a couple layers should that be layers 1 and 2 or or 3 and 4. I'm guessing since you said 2 smallest scales that would be 1 and 2, but just making sure. Also, should I be using default settings for the NR on those 2 layers or will I need to adjust them? Guessing since I'll be using a mask that the defaults will work, but again, just checking. Thanks






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics