Local Normalization help

[jamesRC]

Hello people,

I am trying to use Local Normalization and dont understand something. I have a batch of registered files, two batches from different days in fact. Everything the same except small difference in exposure time, and seeing. I Registered them all into one big list. Then I opened LocalNormalization and loaded all the registered files. The parameters were set to default except I had
"apply Normalization" set to "always" and checked "Generate Normalization Data". Then when I tried to load normalization data, which I can see in the window, I get the message "No local normalization data files have been assigned to integration source images". What is this about, please? What have I missed?

Thank you!

James

[jamesRC]

aha! Ask a question and get my answer , just not where I expected it from. Are the new "abcd.xnml" files I see, the normalized files, so it's all been done? But why is the "load local normalization files" enabled in ImageIntegration? Hmmm not really an answer. Help, please.

James

[pfile]

well, LN has several modes. the default mode is to generate .xnml files only, when run in global mode. that's probably the mode you want, rather than "apply normalization".

the idea is that rather than outputting a whole new copy of the image which have been normalized, it just outputs a file that describes how the file should be normalized. ImageIntegration knows how to read this file and do the normalization internally while it is integrating the frames. the .xnml files are small in size, which is a bonus.

so what you do then is load the _c_r.xisf files (that is, the calibrated and registered files) into ImageIntegration using the "Add Files" button, then, using the "Add L. Norm Files" button, select all the .xnml files. ImageIntegration will match them all up. then you set the Normalization mode in ImageIntegration to "Local Normalization" to make ImageIntegration use the .xnml files. the nice thing here is that because you are using the same input image files, you can easily compare the results of regular normalization and local normalization just by switching the normalization mode in II and re-running.

rob

[jamesRC]

Thank you, pfile.

I thought there was some kind of detail like that wrong. I'm not a super profound user, though I try.

Thank you again! I'm off to try your advice'

James

[pfile]

i'll just add the caveat that using LN is trickier than the tutorials say. you need to use the cleanest reference you can - for me that's a carefully DBE'd integrated master - and you will likely have to at least tweak the scale parameter to avoid dark patches around some structures. IMO it's not something that you just insert in your flow without careful tweaking and evaluation of the results.

rob

[jamesRC]

Thank you, pfile. I tried it as you describe it, on a collection of 210 images over 2 nights at 45 and at 60 second exposure times. My target was the Pinwheel Galaxy, Messier 101. Done as you said, with defaults, it worked. That's great :-). I have tried to include an image of the result but the bands are not visible in the reduced image. It seems to be clearer especially in the background. It doesn't show well in the image but there are what look like extensions of the spiral arms well out from the galaxy, in the same direction. I wonder if these might be LN artifacts?

James

[jamesRC]

Here's a well reduced image. The sky was hazy at the time but some mottling may be due to LN?

James

[pfile]

here is a single 1800s lum exposure of M101 that i took from a dark site with a 12" telescope for comparison. there's no real processing in this image other than stretching so i'm fairly confident what's shown is 'real', and so it should be a good reference.

https://www.astrobin.com/full/186892/0/

rob

[jamesRC]

Hey rob!

My little 120mm apo plus PixInsight, did a pretty good job :-) My image had a rather light background, too. Spent some time reducing that. Now I have an idea what to expect. Thank you very much for the comparison insight :-)  :-)

James