›

DBE problem

Ronald Clanton
in PixInsight
Adam,

I've posted this issue in the PixInsight forum, but haven't received a solution.  After many months of successfully using DBE, I suddenly have been getting weird results with DBE. I've attached a compressed image where you can see the molted/patchy effect. I'm not sure if I'm doing something wrong with DBE or if there something wrong with the underlying image and/or calibration. I've also included a link to the XISF file from WBPP.

Thanks,

Ron

P.S. There is some underlying banding in the image that I have sourced back to my flats.  It seems my LED panel is causing this effect.

Comments

  • BlockHeadBlockHead
    edited February 2023
    What were your DBE setting?
    Screen shot of your samples and settings...

    I have a couple of comments... and I am sorry to bring bad news.

    1. Your data is poor. The gradients are too bright compared to the object you are trying to capture. You simply need a darker sky. Some things might be out of reach in your situation... especially with this sensor and light pollution.
    2. The gradients (sky background) must be varying wildly from image to image. The gradients are very complex...I am assuming your flats are NOT complex in any form. This could be from poorly acquired flats..but I cannot tell. Given how bright the sky is..I would guess these are the conditions you are imaging from.
    3. I suspect your approach to using DBE in this situation may not be optimal. Your image obviously has more than a simple linear gradient. Gradients are additive (in this case). So..if you STEP ONE remove the linear gradient first... then you can STEP TWO flatten the complex components of the "gradients" (light pollution).  

    Regarding #3 you first apply a handful of points around the image and use a large smoothing factor to just get rid of the linear gradient. Then you point samples with a more aggressive smoothing (less smoothing..the default is strong enough) in the high and low areas. It is important to point many points near he boundaries of these regions.

    The attached image is what I get when I apply my technique. 

    -the Blockhead

    Capture.JPG
    1116 x 914 - 338K
  • Ronald Clanton
    Adam,

    Thanks for the analysis!

    So first, I began to suspect that my flats have a problem.  I took a new set this morning using the sky through a white t-shirt (no LED panel) and reran WBPP.  Much better!  I still have some molting, which I think may be coming from the t-shirt.  I ordered a new Aurora panel and hope that will fix that problem.

    I used your technique and the results are significantly better.  I still can't match your results... but oh well.

    Finally, I'm questioning my acquisition of the lights.  I'm using a ASI533MC Pro (OSC) with my HyperStar (F1.9).  I took 240 60-second images.  I talked to Starizona and they recommend that I take 25-40 second exposures.  So that may explain the washed out look.  Plus, I live in Bortle 5 skies... but the seeing was below average.... and the moon is still somewhat of a factor.

    Historically I've had good luck with my 0.7 reducer on my Edge 8HD... but the HyperStar is a whole new learning experience for me.  Additionally, I need to invest a bit more resources in getting better flats.

    Thanks again!

    Ron


  • BlockHeadBlockHead
    The moon!!!!
    Yes..this would contribute quite a bit. The T-shirts are also an issue.
    You have number of avenues to pursue to minimize this issue.
    May I use your image for a quick demonstration of DBE on my site?

    -the Blockhead
  • Ronald Clanton
    Yes... that stinking moon!  LOL!

    Yes, you may use the image any way you like.

    Thanks,

    Ron
  • Peter Rejto
    Ron,

    In case you are interested you might want to check out this post:


    Peter
  • Ronald Clanton
    Peter,

    Yes, that is what pushed me to order the Aurora light panel.

    Also, I just watched Adam's YouTube video using my data.  I'm glad that I could provide this community service to the group... lol.  What an awesome job by Adam!  I'm really impressed with the power of using SXT and DBE to solve these kinds of problems.  It's great that Adam is around to figure out these things for us.

    I think I'm going to have to bump my membership to the Horizon level.

    Cheers!

    Ron

  • Patrick Bisaillon
    @BlockHead Ok I admit it, I felt a bit dumb when I watched this video and heard your comment about the use of t-shirts in the process of shooting flats haha but many people do this so I didn't feel so bad afterall :)

    My question is: for my widefield FRA300 scope (or any other similar scopes in size), what would you recommend to replace the t-shirt technology?

    Thank you  
  • Phil MisiowiecPhil Misiowiec
    Deep Sky Dad makes great flat panels - I have two of them. Basically he 3D prints the assembly and adds a light panel and servo motor. The ASCOM driver is on the product pages. Takes a couple months to receive one after you place an order. He's based in Slovenia.

    They work great with NINA. I have an advanced sequence that captures flats at the end of each imaging session and then shuts down the entire system (mount and mini PC). 

Sign In or Register to comment.