PixInsight NormalizeScaleGradient Part 1: Introduction

[@nightskyimaging]

AWESOME EXPLANATION! Thank you for doing this.....I am now ready to revisit prior data and redo! Now I have plenty to do on these cloudy nights!

[@AdamBlock]

I do hope others reproduce my experiment and compare the weights they get ..and see that NSG is doing it properly where the default in iMageINtegration is not. These are early days..and I might be missing something- but so far I have not seen an issue.

[@claude77573]

Great video and great script! The author said this was 27,000 lines of code! Holy Cow! I will definitely make a donation for a "cup of coffee"! I wonder if the script can be tuned to hammer the daylights out of available computer resources. I have a 16 core, 32 thread Threadripper system, and the script uses only about 10% of my system's capacity. If it can be tuned to use all available resources (like other scripts and processes), then this computationally intensive process would really rip!

[@neverfox]

So one thing that this seems to give up, if I understand correctly, is weighting based on things other than noise, e.g. FWHM and/or Eccentricity. This is an improvement on the standard noise evaluation but when using something like WBPPWGHT, standard noise evaluation isn't being used anyway. Would you recommend going entirely with NWEIGHT or working it into a compound weight with other factors like FWHM?

[@AdamBlock]

Ok... so your question reveals something about my experience and perspective. I have really never been a fan of weighting images based on FWHM or the other metrics. Too much emphasis on this! Usually the quality of an image is based on the ultimate S/N... and when you give crappy images more weight than they are worth you are hurting this ultimate result. So... I value this signal-noise based weighting over things found in SFS. Good question...

[@mikeschi9551]

Great video set, explanation and demonstration! Part 3 really brings the power of NSG home - thank you. Quick question; I assume we'll still use DBE or ABE - correct?

[@AdamBlock]

Yes...absolutely. Normalization is simply an image "matching" process that eventually results in images compatible for rejection algorithms and weighting. But the gradients, matched to a reference, are still there. You need to take care of them.

[@yomichee]

Invaluable content thank you! Btw are your course fees in Canadian dollars ? Tia

[@AdamBlock]

Thank you for watching!! No... it is USD. However, the card processor takes care of the exchange rate I think.

[@trevorgreen2232]

I hope this is not a dumb question but do you use the same reference image for all colours and lum ?

[@AdamBlock]

A VERY GOOD QUESTION. i should have repeated this... you need a reference for each color... do not use a Red reference for your Green data!!

[@subashisamarasinghe1439]

Hi Adam, this script is not in batch processing section of scripts , I am using pixinsight 1.8.9-1. Is there a way to install the script?

[@AdamBlock]

Yeah... politics. see www.youtube.com/@NormalizeScaleGradient

[@dkuchta5]

How does this handle images where most or all of the "background" is actually part of the nebula you're imaging? DBE or ABE will tend to remove some of the nebula because it thinks it is background gradient. Will this do the same?

[@AdamBlock]

NO. Normalization is a frame matching calculation. You are comparing the reference frame to the target frame... and at each pixel (or region of pixels) the SAME FLUX (light, data..etc) should be there. The difference between the frames is what matters... and the gradient is part of this. It doesn't matter if there is nebulosity or not filling the entire image- all of the math works out. This is different than trying to taken samples and estimate the background (without comparing to other images...)

[@dkuchta5]

@@AdamBlock Excellent! Thanks for the reply.

[@neverfox]

So it seems that you can't use drizzle with this technique? It won't allow me to add drizzle data from the registration step if I load nsg files into ImageIntegration.

[@neverfox]

Seems that it was just an issue of the filenames needing to match (which is a bit annoying), but it raises the question: will drizzle integration benefit from the NSG weighting?

[@AdamBlock]

Yes you are correct. Since the images need to be aligned (in the normal way) to do the magic measurements... no drizzle. The developer is considering this problem... so perhaps in a future version.

[@blueridgedsia]

If you shot 2 groups of images at 600s and 300s, could you still load both groups into image integration and integrate them into one image?

[@AdamBlock]

At the moment you will get two different integrated results. I have talked to the developer about making this possible. It does add another level of complexity (that often confuses people). In addition, with the release of NormalizeScaleGradient- you will probably *not* want to ImageIntegration to normalize your images (which is what happens). I recommend watching my videos on NSG.

[@blueridgedsia]

@@AdamBlock Thanks

[@kayedsss]

Thank you very much