›

NSG on different nights data (same object, very different quality)

MAURIZIO MOLLICA
in PixInsight
Dear Adam,

please consider this situation:

You have two sets of data on the same object but taken from different  sky (eg. set 1 from bortle 4, set 2 from Bortle 5/6 nearby a sodium street lamp).

Now you have properly calibrated and registered all the lights.

Well, if you are going to use NSG, putting all the data togheter, you will end up with the entire second set eliminated due to the weight being too low.

That being said, what would Adam Block do in this case, considering I would not want to eliminate the second set?

Whath would be a good compromise, in your opinion?

Please look at the attached files.

Thanks in advance,

Maurizio

Comments

  • BlockHeadBlockHead
    Well... there are two points.

    1. The data with the low weights are crappy data (assuming the exposure time is the time)... at least relative to the other data you acquired. There is no way to sugarcoat this and the numbers are not lying. This is a quantitative measurement done by measuring stars (photometry). 

    2. You certainly can include this data for what it is worth. Right it is being excluded only because of the default threshold. You are welcome to lower the threshold. Then the only thing that will happen is that the data will be included into the stack for at the calculated weights. However, this does change the fact the poorer data will not contribute as much.

    So here is what "Adam Block" would do. I would first integrate (stack) the data with only the good frames. Then I would stack the data including the poorer frames. I would compare the two masterlights. If including the poorer data makes a difference... then you have your answer! However, I am going to guess that its inclusion may not be helpful. This is because the extra signal of the poorer data perahps comes along with other contaminating issues like more gradients and things.

    -the Blockhead
  • MAURIZIO MOLLICA
    Thank you Adam, 

    I did exactly what you said in point 2 and, guess what, there were no appreciable differences compared to 

    the first stack...

    In order to not exclude the second set I had to lower the threshold a lot, hence their contribution was 

    very, very low...

    Lesson learned!

    Thank you!
Sign In or Register to comment.