›

M45 color

auroraTDunn
in PixInsight
So how do I bring out all the crisp color in M45 like Adams image?
I have almost 30 hours of LRGB data using a RedCat71 and an ZWO ASI6200MM.
I've stacked the data in WBPP and have tried many workflows to bring out the colors but I keep getting nice blues but everything else is muddy.

My basic process has been, for this data:
RGB combination
DBE or GraXpert (I actually like my GraXpert results on my data, no symmetric issues like on Adams DBE tutorial)
SPCC
SCNR

The next bit I tried several combination but am liking the following to get to nonlinear star and starless images and not my normal order
A few small histogramTransforms
SXT
NXT
I then use curves and a couple other things on the starless image
I then used Adam's trick of a second background extraction like Adam demos in his NGC1333 WORKFLOW (this brought out some GREAT detail!)

For my Lum I kinda do the same as above but then add some UnsharpMask processing and a bit more work to balance the contrast nicely for combining.
I then do an LRGB of the Lum onto my RGB
I also used the Lum as a Lum mask to work on cleaning up the background a bit as well.

Then back to Curves and into Photoshop for some RAW processing (yes I have done a few other things for the overall image processing such as an extra round of NXT, BXT before Photoshop, ect, but I'm just trying to concentrate on the color processing right now)

I have not bothered working on my stars yet as I'm still not happy with the lack of color other then my blues.
I've tried a couple quick stabs at masking the blue channel and then working on the rest but again I still just getting muddy colors, nothing vibrant like I can see in Adam's and a couple other images.

I went back and tried drizzling 2x during WBPP, but it bought me close to nothing except for the smallest stars. But maybe I'm wrong in not using drizzled data?

Should I, somewhere in here, divide the image back out into seperate R,G,B channels and then process the Red and maybe tone down the green a hair (worried I'd start turning my blues non-blue if dinking with the green channel to much) then recombine them? Never tried that but???

Whats sad is I'm a color expert, scientific visualization wise (as that's my career)  and I keep failing and thus its hard for me to finally quit my tunnel vision processing approach and ask for help. But alas PLEASE!


Comments

  • auroraTDunn
    Sorry folks I failed to attached the 4 master files (L,R,G,B). These are not drizzled.

    Note the file size is 730MB for the four files ( I did mention I use a ASI6200mm - big files)
  • BlockHeadBlockHead
    Sorry...I am a little confused. Why are you uploading the LN references and not the master lights?

    -the Blockhead
  • BlockHeadBlockHead
    Tim,

    Even though you gave me low quality data since it is not the entire stacked image and only the LN images... I think you make my point.  I just recently asked if I could announce my new course on Comet image processing on the FB group called "Comet Watch" .. .the answer No.. the first reason given- Hey Adam, "These are seasoned expert imagers on this group... "  OK Fine. 

    It isn't the equipment and site- yes, they help. But the decisions in processing based on how things work as well. Some things are simply not known or appreciated. You gave me an image... see the attached. I labeled the colors in the image. The data is there. I don't know the quality of your data in the stacked image- but that is just an acquisition issue. What is in my image is in your data. So it isn't just blue. The color are there and the game is to get enough clean images to bring out (increase the contrast) of this information.

    I think you should drop any inclusion of luminance for a moment. A basic principle of Luminance blending... *it does not help color information* . You are fighting blending luminance to retain color. Another principle- Unsharp mask- likely isn't going to help you either. This is about COLOR not detail, not brightness. 

    Graxpert? Not on this field. I feel that is a poor choice. How do you know you have an optimal result? I can demonstrate on fields that are filled with important color information throughout the field that Graxpert fails. I do not think your data has an issue like my original tutorial presents- it is easier now that stars can be removed.

    -the Blockhead

    Screenshot 2024-02-11 210219.jpg
    1908 x 1263 - 224K
  • BlockHeadBlockHead
    edited February 2024
    I just re-read my message.
    Let me state this a little more clearly.
    When I began astrophotography and generating images 25 years ago- there were no references.
    There was nothing to compare to in terms of other RGB images of many objects. I learned to look at the data alone to make the decisions and choices that I did. This did not mean I could just fiddle until I got what I wanted..it also means I needed to appreciate what was significant information in an image and not noise or artificial (artifacts). This took some practice. 

    Notice my first attempt was not as good as later adjustments. See this image:

    Note how I darked the sky, raised the color saturation and (I think) did a large scale USM to get the latest result. But the original has everything there. It just wasn't quite final I think- but even this attempt was STILL my confident rendering of this data in a way most people (any?) did not... because I think they kept looking at other people's images. 
  • auroraTDunn
    Thanks Adam,
    And sorry about the wrong images, I'm a moron, I spend Sundays working on personal artwork of a totally different nature, realized I never posted the, so quickly opened the folder and just grabbed the top 4 files without even paying attention to which I actually grabbed.

    And thanks for the comments. Late last night I came up with an idea to do some funky mask scripting to get just the areas I want and if I still can;t refine it enough I'll break down and do what I do best, manhandle them in Photoshop and then bring them back into PI, or maybe just work with it all in PS till I'm ready to pop the stars back in. The plan was always to do my final processing in PS anyways. I was just hoping for a PI solution. Just need the ability to paint your masks in PI and life would be magic!

    As for the comet series - I saw the intro video last night and got all excited. I spent an eternity on the couple of comet images I liked enough to keep working on so barely wait to go through this new series to improve my old images and as prep, I hope, for new ones this year!


    The link to the actual files are here for anyone that wants to see my data (the right ones this time!)
Sign In or Register to comment.