›

Two subjects, Under sampled images and DBE on highly nebulous images.

Kevin Fordham
in PixInsight
Hi Adam,
I have just finished my first run through of the Fundamentals video series and I have learned so much. I am trying to transfer the Galaxy processing to Nebula images that I have done that are mostly in Narrowband.
I have noticed that my stars are significantly smaller (pixel size) than some of your examples and have used the Drizzle integration process to good effect and it has produced larger stars. I just wondered if you have any thoughts about under sampled images. I have been using a Starwave 70 Quad with and ZWO1600 mono for a nice wide field, but this setup does seem to produce the under sampled images I refer too.
The other question I have relates to the Dynamic DBE process when used on highly nebulous images such as NGC 2244 or even NGC 7000 that is full of nebulous data.
I am desperately trying to get to grips with processing but it is still the most frustrating part of this fascinating pastime, so any pointers you can give would be most appreciated.
Keep up the great work.
Kevin

Comments

  • BlockHeadBlockHead
    Hi Kevin,

    Yeah... your stars will be smaller! The data I have been typically demonstrating with is captured at 0.33"/pixel. If the seeing was 2"... these are 6 pixel stars!  However, I have also been doing some work with a 180mm Tak (Epsilon astrograph). Here the plate scale is 5"/pixel. The stars are always undersampled:


    However, when I combine many frames the combination of dithering the images plus the bicubic resampling (likely what the Auto method chooses)  that occurs during registration produces round stars in a combined image. So if the issue is just getting round stars, I haven't felt the need to worry about drizzling data. However if you want to zoom into your undersampled data to see small/fine detail- then using drizzling to recover the lost spatial information along the diagonals makes sense. In both the natural case I mention without drizzling or when using this option- it is important to have lots and lots of frames (more than 15 I would say.

    Here is the thing- if the nebula is smooth to begin with and the stars end up round without drizzle- I don't see a benefit. If however, you have boxy stars and fine detail to see when you zoom in- then drizzle is a good idea- but only if you have a lot of frames. I believe drizzling does introduce some noise.

    I should probably make a video of combining the data from the image above to demonstrate the drizzle integration. 

    Concerning DBE on a nebulous field- the real question is can you tell there are gradients to correct? If you cannot distinguish the gradients from real features- there really isn't a path forward. You *can* do a prescriptive form of DBE. Let us say that in your previous images you *always* have vignetting in the corners. So you could apply a points near the corners on bits of sky that appear to be nearly sky values and then use circular symmetry to model the vignetting.

    Now if you *can* tell there are gradients... what does it look like? In the case of NGC 7000 there are many dark bits of nebulosity. These dark portions should all be somewhat neutral in color (for example). If some of the dark bits are blue in one part of the nebula but greenish in another part... then you can safely assume you are seeing the color gradients. The samples would then be placed on the dark nebula bits to be your best sky values. 

    A combination of the two techniques I mentioned above will likely take care of any major gradients and render a faithful color image. 

    -the Blockhead
  • Kevin Fordham
    Hi Adam,
    It’s funny, I have been watching your videos so much lately that when I just read your reply I could hear your voice in my head, I need to get out more!
    Thank you for response, one of the things I am getting from the Fundimentals videos is really look at the data you are working with and don’t be afraid to experiment a little and have an idea what you are trying to achieve with the process you are using.
    Thanks,
    Kevin
  • BlockHeadBlockHead
    Kevin, 

    Just a heads up...I have completed the section on Drizzle. I will be releasing the OSC lessons first... then Drizzle. So... it is on the way!

    -the Blockhead
  • Kevin Fordham
    Thats great, Thanks for the info.
  • BlockHeadBlockHead
    OSC has been released...

Sign In or Register to comment.