›

How do I make my images more realistic and less 'colorized' looking?

Greg Erianne
in General
Hi Adam,

I've been working with PixInsight for almost a year and I'm becoming familiar with how to use the tools, although I admittedly have a long way to go to feel very competent.

It seems, though, that I can never get realistic looking images in the sense that mine always seem 'colorized' and artificial.  To me, the best images are clear, aren't overly saturated or colored, and (this is the thing that eludes me) they seem almost transparent.

Are there particular videos on your website that might address my problem?   Is there anything else you can suggest as a start?

Thanks so much, and thank you as always for all the hard work you put into your work, your videos, and your fantastic web site!

Best,
Greg

Comments

  • BlockHeadBlockHead
    Hi Greg,

    Please begin by showing me a picture that shows your work..and another that shows what you are striving to achieve. If you want to brown-nose... choose one of my pictures that fits the bill. :)

    -the Blockhead 
  • Greg Erianne
    Thanks so much for replying, Adam.

    I'm sorry the description of my problem is so 'nebulous', but being a left-brain dominant person and having little innate artistic ability, it's very tough for me to explain my subjective feelings when it comes to art/photography.  I'm working on it, though!

    To start, I attached two images of the Jellyfish Nebula.  The second one is mine.  For comparison, I've attached a JPG (first one) from a member of the Society to which I belong. If you notice, the Jellyfish in his image is sharp, and 'wispy', rather than looking solid and colored-in.  The colors and brightnesses are different, of course, but I'm really interested in the substance of the Jellyfish part of the image.  Achieving that wispy, ethereal look just completely escapes me it seems.

    I also attached an image of the Rosette Nebula (from a FB member) that seems to me to be perfect in the sense that it looks almost like you could step into it.  Yours also has that same 'authentic' look.  Mine (the last one) on the other hand (although it's an older image) just looks like the colors were pasted on.  

    Sorry for the lengthy post, but I wanted to try and explain a bit better.   I hope all that makes some sense to you.

    Thank you again!

    Greg
    jellyfishFB2 (2).jpg
    2678 x 1991 - 3M
    IC433_Final_GreenBkgdRem2.jpg
    6248 x 4176 - 32M
    Lindas_Rosette_Nebula.jpg
    1695 x 960 - 247K
    AdamBlock_rosetteteles_orig.jpg
    675 x 900 - 282K
    MyRosette.jpg
    6245 x 4172 - 7M
  • BlockHeadBlockHead
    There are two issues you are encountering. You need to understand the issues first- then the solutions will make more sense. In the case of the Jellyfish- you need to stretch the image more- but not make either the color or brightest parts become "solid" (equal in brightness). Look at the first image closely... do you see how the nebula is "filled in"? Your image has a darkness in the middle of the nebula... not in the first image. You need to consider not have a black background. This is harder (surprised right?) because this will reveal all of the issues that a clipped sky hides. 

    Do the following experiment... take your image, using ImageBlend or PixelMath, and screen blend it with itself. Do you see any attributes of this version you like? What if you took your original linear image of the JellyFish and did a masked stretch? What about exponential transformation?

    Do you note the items I am mentioning?
    Screen Blend
    Masked Stretch
    Exponential
    (GHS should be here as well... but the above show an effect in a single button press)

    None of the above will instantly give you the first images appearance. But you are looking for an incremental step. Can you honestly tell me that none of these suggestions is a move in a Direction you want to go? If you can say that it is a step in the right direction... then *this* is how you approach processing. You find a process or technique that moves in the direction you want... it will never fully give you everything... there will be attributes of any process or technique that will also be counter to something you want. But the net gain is what you are looking for.

    Regarding your Rosette (narrowband)...that is a DIFFERENT problem. I hope you can see that. One data set (color and brightness) is dominating (the OIII). 

    The techniques of NB normalization (the script) or the narrowband modifier methods will deal with this. I demonstrate both in Horizons in the narrowband videos. I have yet to...but need to... make a video that permits this with the new NB ColourMapper script. It will make it easy to do.

    -the Blockhead 
  • Greg Erianne
    Thank you, Adam, for taking the time to write a lengthy reply and highlight the problems I'm having.

    I have a lot to analyze here, and a lot to try.  This is a fantastic starting point, which is exactly what I was hoping for.  

    Your help and advice is greatly appreciated!

    Best,
    Greg
Sign In or Register to comment.