›

Grid pattern in my combined RGB image...

Daniel Renner
in PixInsight
hi! I have this very strange grid pattern in my combined RGB image... Im shooting mono (asi 1600mm pro) and I stack in the latest version of Wbpp. The individual files look fine, but it is when I combine the image and doing noisereduction that the grid really shows.
Does anyone know what is wrong?

(I dither.
I have not enabled CFA.)

Comments

  • BlockHeadBlockHead
    I am curious. What normalization method did you use?
    -the BLockhead
  • Daniel Renner
    I used everything as default with maximum quality preset. 
    1.PNG
    1306 x 830 - 138K
    2.PNG
    435 x 297 - 10K
  • BlockHeadBlockHead
    edited March 2022
    First...there are no defaults... just need to make that clear. 

    I would suggest integrating the images without Local Normalization to see if there is a difference.
    LN uses image partitions at a particular size (squares). I wonder if you are seeing enhancements due to tiny errors in LN. This would result in a grid like result I would think.

    As another test, I would use NSG (if you have it) and see if you get the same result. If you do not get the issue with NSG (which uses a different method)- then it points to the normalization as being the potential issue.

    I a speculating, but I think it is good thing to check out  .... but let me ask you- has anyone suggested this as a source of the problem yet?
    This is the kind of thing that I hope people find value in with respect to my efforts with PixInsight.

    -the Blockhead
  • Daniel Renner
    Yes I forgot to mention that when I stack in NSG I dont get that grid at all , and I had LN enabled. Its something with the integration in wbpp that does something wierd... 

    Ill try the integration without LN to see if it gets better, but then NSG is needed anyway...

    Thanks for helping me out=)

    If you want to try this for a video I can send you the files, let me know.
  • BlockHeadBlockHead
    Not exactly... you are just using additive+scaling (I think) when you disable LN in WBPP. 
    You are still normalizing the data.. it just isn't local..it is the local part that I think could be the issue.
    I think this strongly points to LN. 

    Have you posted this example on the official forum?

    Yes, this could make a very nice video... 

    -the Blockhead
  • BlockHeadBlockHead
    Also, could you please post the result with NSG that I assume looks cleaner?

    -the Blockhead
  • Daniel Renner
    hmm I didnt save the grid image so Im trying to remake it, but now it looks normal... WIERD!
    I dont know what was wrong before. But now it looks fine...
    If I get the grid pattern again ill find out more detailed what was wrong and Ill post here again.

    /Daniel

  • Daniel Renner
    Okey! now it has this grid pattern even with NSG! What I did was: MURE denoise, crop, DBE och each channel, RGB combine, then DBE again and now it has this grid. I know its super stretched. Its not as noticible when its correctly stretched, but still, its wierd..
    nsg.PNG
    1452 x 1083 - 3M
  • Daniel Renner
    edited March 2022
    I have prepared all darks, flats, darkflats, and lights on google drive ready for you to look at IF you want, let me know and Ill send you the link on facebook =)


  • Daniel Renner
    I also noticed some kind of grid on my master flat, I dont know if its normal or if it could come from my Tshirt when doing flats? 

    (image has been stretched in a way to see the grid)
    3.PNG
    1582 x 1204 - 2M
  • BlockHeadBlockHead
    Wow..that could certainly do it...BUT, you would see this pattern in all of your results no matter than method you used to combine the data...it would persist.But you may be on to something. 
  • Daniel Renner
    hmm I could try to retake my flats without a tshirt to see if that really is the problem... Ill do that tomorrow and update you later =)
  • BlockHeadBlockHead
    excellent
Sign In or Register to comment.