NormalizeScaleGradient Script

[John Murphy]

Adam Block has just made a Horizon video tutorial on how to use the new NSG features.

https://www.adamblockstudios.com/

[Mark de Regt]

Thanks, John!

[rockenrock]

Hi John,

I first used NSG at Ver 0.8. Since then there have been many updates and improvements. Thanks!

 

1. I have not reprocessed any previously run NSG data with latest versions of NSG as the days and months have gone by.

Would you kindly think back to which older version of NSG that before which I should reprocess my subs with NSG2.1.4? 

 

2. I really like the new Correct Target checkbox under the Gradient Graph expansion. 

I would like to make this corrected image more contrasty so can see the result better.  Anything for that?

 

3. For me, I have a lot of light pollution and stray light coming in, and finding the best reference image is important. When I screen stretch some subs I  may think it has a simple gradient, but the inherent gradient is hiding behind the light pollution, stray light, etc. 

Then what looks to be a good reference image (my guess) is not showing up so good after NSG and integration. I am willing to spend more time to find the best image.  Is there way I could check different potential reference images by inserting a false 'perfect image' as reference, then checking to find which real sub has the most acceptable Gradient Graph. Then use that real sub as my reference?

 

Many thanks!

    Roger

 

[John Murphy]

Hi Roger

I have been putting together some release notes:

 

NSG 2.1.4 (18th March 2022)
===========================
* Revert menu name from 'NormalizeScaleGradient 2.0' back to 'NormalizeScaleGradient'. The '2.0' was preventing the Reference Documentation from displaying.

* Fixed 'Reset' bug. The Reference image text field was not being reset.

* If running on the previous PixInsight 1.8.8-12 version, removed an unnecessary warning asking the user to create an ImageIntegration process icon.

 

NSG 2.1.0 (3rd March 2022)
==========================
* Gradient dialog: Added ability to blink between the reference and corrected target image. What you see is what you get! This makes it much easier to determine the correct amount of smoothness to apply. It is now apparent that much less smoothing can produce great results.

* Changed 'Auto exit' to 'Auto exit and run ImageIntegration'.

 

NSG 2.0.2 (24th Feb 2022)
==========================
* I have rewritten the code I was using to read the FITS headers. The code is now both robust and fast. This should significantly reduce the time required to load images.

* The gradient calculation should also run more quickly.

 

NSG 2.0.1 (18th Feb 2022)
==========================
* Drizzle data files are now added to ImageIntegration.
* Improved file I/O error handling
* Updated Reference Documentation

NSGXnml: 
Updated to remove the limit on file path length of 225 characters to the maximum allowed by the operating system

 

NSG 1.6.1 (6th Feb 2022)
=========================
* Improved gradient correction.
* Images opened by double clicking on a target list image now only have an STF applied. Previously it was stretched, which may have been confusing.
* If focal length or pixel size is missing from the FITS header, an attempt is now made to read them from the view's properties.
* Gradient images are now full size.
* If there is no filter name, this is now displayed as "" instead of "NONE"
* FITS Headers are now read from the PixInsight 'FileFormatInstance' instead of a JavaScrpt method. In rare cases, the JavaScript method could fail to find all headers.

 

NSG 1.6 (24th Jan 2022)
========================
* Improved error handling. In the Output Images section, I have added an On error: combo box, with options: Continue, Abort, Ask user. This fixes an issue where if all the images fail, it was previously necessary to click an 'OK' dialog once for every file.
* Full paths toggle button. This allows the target image filename to be displayed with or without its full path. The Filename column's tool tip always displays the full path.
* Extra columns in the Target Images table. All columns are sortable by clicking on the column header:
PSF Weight This displays the PSF Signal Weight, normalized to the 0.0 to 1.0 range. Higher is better.
Noise This displays a NOISExx noise estimate, scaled by the exposure time and airmass (NSG has not calculated the brightness scale yet). It is normalized to the 0.0 to 1.0 range. Lower is better.
Airmass If this is available in the FITS Header, it will be displayed.
Time Exposure time in seconds. This is only for display. NSG will scale images correctly without needing to know the exposure time.
Filter The filter must match the reference filter.
* Table double click Double click on a Table row to display the image.
* If the altitude is less than 10, the number is prefixed with '0' to ensure the numbers are sorted correctly.

 

NSG 1.5 (9th Jan 2022)
=======================
* Images are rescaled to avoid truncating high values.
* 'Rescale result' checkbox has been added.
* The reference image text field is cleared if it is no longer in the target images list.
* Bug fix: If the input file was in Integer format, the corrected file would lose all its original FITS headers. This has now been fixed.
* Bug fix: If the image was severely black clipped (entirely zero / black background), the script could fail. This has now been fixed.

 

Regards, John

[John Murphy]

I think the most important improvement was the ability to show the corrected target image. The reference checkbox can then be used to blink between the reference image and the corrected target. What you see is what you get! This better feedback allows the user to select the best smoothness setting. It turns out that much lower levels of smoothing can produce extremely accurate results without introducing artifacts. Smoothness levels of -2.0 or -3.0 can produce excellent results.

 

Having the ability to modify the stretch is an interesting idea. I will consider adding this in a future version.

 

If you have an 'ideal' image, provided it is linear, you can use this as the reference. This should help you remove all gradients. The corrected target images will still reflect your data, just without the gradients.

 

Regards, John Murphy

[Giulio]

On 3/19/2022 at 2:47 AM, John Murphy said:

PI 1.8.9 changed the '.xnml' data file format. The files are not backward compatible. Unfortunately, this means that if you created one or more of the '.xnml' files before 1.8.9, they cannot be used within 1.8.9 

 

I had to make a minor change to the NSGXnml C++ module (the API had changed as well). It is important that the correct NSGXnml version is installed. If this error occurred with data that you created in 1.8.9, I would recommend uninstalling NSGXnml, and reinstalling the PI 1.8.9 NSGXnml version.

 

To uninstall NSGXnml:

• PROCESS > Modules > Manage Modules...
• Scroll down to the NSGXnml entry and deselect it.
• Press 'Done'.

To install it:

• Unpack the compressed .zip / .tar.xz file.
• Copy NSGXnml-pxm.dll (windows) / NSGXnml-pxm.dylib (mac) / NSGXnml-pxm.so (linux) to any folder (but don't use a PixInsight folder)
• In the PixInsight PROCESS menu, select:
• PROCESS > Modules > Install Modules...
• Browse to the folder. Note that the folder contents are not displayed.
• Press 'Search'. A dialog should display '1 additional PixInsight module(s) were found...
• Press 'Install'

If you are installing on a Mac, See the Extra note for Mac OS on the https://sites.google.com/view/normalizescalegradient website before installing.

Hi John,

I followed this procedure to install the PI 1.8.9 NSGXnml version, because of the issue with xnml files; I press search on the folder where is the file NSGXnml-pxm.so but  Install Modules do not find it. I run Pixinsight on Ubuntu. 

I did the same in Windows 10 and here I was able to upgrade to the PI 1.8.9 NSGXnml version without problem.

Giulio

Edited April 1, 2022 by Giulio

[John Murphy]

4 hours ago, Giulio said:

I followed this procedure to install the PI 1.8.9 NSGXnml version, because of the issue with xnml files; I press search on the folder where is the file NSGXnml-pxm.so but  Install Modules do not find it. I run Pixinsight on Ubuntu. 

I did the same in Windows 10 and here I was able to upgrade to the PI 1.8.9 NSGXnml version without problem.

I would first try to install using the repository. You can find details on my website:

https://sites.google.com/view/normalizescalegradient

 

However, I think this may be a PixInsight - Ubuntu problem. Another user had the same problem with Ubuntu, and saw the error message:

*** PixInsight API Error: Module: /opt/PixInsight/bin/NSGXnml-pxm.so

Cannot load library /opt/PixInsight/bin/NSGXnml-pxm.so: (/lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /opt/PixInsight/bin/NSGXnml-pxm.so)): Module load error

Unfortunately there is not much I can do about this. If this is the cause of your problem too, and you are knowledgeable about Linux, you might be able to resolve this GLIBC_2.33 problem.

 

I believe that StarNet V2 may have similar problems on Ubuntu:

https://pixinsight.com/forum/index.php?threads/starnet2-installation-in-ubuntu.17999/

 

Sorry I can't provide more help, John Murphy

Edited April 1, 2022 by John Murphy

[Giulio]

13 hours ago, John Murphy said:

I would first try to install using the repository. You can find details on my website:

https://sites.google.com/view/normalizescalegradient

 

However, I think this may be a PixInsight - Ubuntu problem. Another user had the same problem with Ubuntu, and saw the error message:

*** PixInsight API Error: Module: /opt/PixInsight/bin/NSGXnml-pxm.so

Cannot load library /opt/PixInsight/bin/NSGXnml-pxm.so: (/lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by /opt/PixInsight/bin/NSGXnml-pxm.so)): Module load error

Unfortunately there is not much I can do about this. If this is the cause of your problem too, and you are knowledgeable about Linux, you might be able to resolve this GLIBC_2.33 problem.

 

I believe that StarNet V2 may have similar problems on Ubuntu:

https://pixinsight.com/forum/index.php?threads/starnet2-installation-in-ubuntu.17999/

 

Sorry I can't provide more help, John Murphy

Thanks for your reply John. I solved! Just select "recursive" next to search in Install Modules.

[Giulio]

Sorry John,I made a mistake. It is not solved. I was not using PixInsight1.8.9_NSGXnml-pxm_linux, but the old one. 

Cheers

 

Giulio

Edited April 2, 2022 by Giulio

[John Murphy]

2 hours ago, Giulio said:

Sorry John,I made a mistake. It is not solved. I was not using PixInsight1.8.9_NSGXnml-pxm_linux, but the old one. 

I will look into this over the next few days.

Regards, John Murphy

[Neverfox]

Update: same mistake; accidentally reinstalled an old version of NSGnml

I'm having the same issue as OhmEye.

*** Error: Incompatible local normalization data version. Expected >= 1, got 0

 

Running NSG 2.1.4 installed from the repository in PI 1.8.9 Mac OS. Preprocessed with WBPP.

Edited April 2, 2022 by Neverfox

[John Murphy]

3 hours ago, Neverfox said:

Update: same mistake; accidentally reinstalled an old version of NSGnml

I'm having the same issue as OhmEye.

*** Error: Incompatible local normalization data version. Expected >= 1, got 0

 

Running NSG 2.1.4 installed from the repository in PI 1.8.9 Mac OS. Preprocessed with WBPP.

To check you have installed the correct NSGXnml version: PROCESS > Modules > Manage Modules...

and then select NSGXnml. The lower half of the dialog will display details about NSGXnml. The critical one is the API version:

The API Version should be:

• PixInsight 1.8.8-12: API version .... 0x172
• PixInsight 1.8.9: API version .... 0x173

If this is not correct, you will get the incompatible error.

On Mac OS, the repository does not contain NSGXnml, so you have to install the correct version manually. This is describe on the NormalizeScaleGradient website.

 

Important

You must first uninstall NSGXnml from PixInsight before attempting to install the new one:

• PROCESS > Modules > Manage Modules...
• Scroll down to the NSGXnml entry and deselect it.
• Press 'Done'.
• Restart PixInsight

After restarting PixInsight, follow the directions on the NSG website.

https://sites.google.com/view/normalizescalegradient

 

I plan to upload some updates tomorrow:

NSG 2.1.5

Reference Documentation 2.1.5

NSGXnml 1.0.3

 

Regards, John Murphy

Edited April 2, 2022 by John Murphy

[rockenrock]

On 3/31/2022 at 3:07 PM, rockenrock said:

Is there way I could check different potential reference images by inserting a false 'perfect image' as reference, then checking to find which real sub has the most acceptable Gradient Graph. Then use that real sub as my reference?

Hi John,

I had asked the above question and you answered: 

"If you have an 'ideal' image, provided it is linear, you can use this as the reference. This should help you remove all gradients. The corrected target images will still reflect your data, just without the gradients."

 

So I created an 'ideal' red channel image using pixel math where I substituted a fixed value for the entire background using this expression: iif($T<.245,0.245,$T). I applied this to one of my (calibrated & registered) subs.  I selected .245 using readout mode and finding a value that was below the object of interest (M106) and above (or equal to) the highest background level.

I used this one sub as my reference image and ran NSG. 

 

After running NSG, the NSG subs all had no complex gradients which plagued my original data. I then integrated my NSG subs and applied a normal screen stretch. No DBE run. See attached comparison.

I am very happy, because previously I spent several hours, and several times applying DBE to get almost as good as this image. The new image looks noisy, but this is because the automatic screen stretch is more aggressive on the more controlled new image.

 

I think I was too aggressive with gradient smoothness (set to -1.8) resulting in over correction near the bright star at top and the galaxies. Do you agree? I am going to try a little lower value.

 

So what do you think of my pixel math technique to create the 'ideal' reference image?

 

Thanks,

     Roger

 

 

 

[Andrew Bicos]

Hi, I just started using NSG v2.1.5 and NSGXnml but I'm having a problem that I hope is easy to solve.  However, I have spent several hours trying, so decided to ask for help.  I opened the script, add my calibrated registered images (ending in _c_cc_r from WBPP), selected a reference image that has a high altitude, entered an output directory and along with the defaults checked Gradient images, checked ImageIntegration and entered a Process icon from the ImageIntegration process using the new instance triangle onto my workspace as a placeholder and checked Auto exit and run ImageIntegration.  I kept all other defaults.  I ran this but noticed that when I opened up the Process icon on my workspace that the fields had been populated but for Normalization it had selected Local normalization, which I knew was wrong since it should be No normalization.  I then checked my output directory and saw that there were no new images with the _nsg appendage - there were only gradient images and xnml files only.  What am I doing wrong?  Thanks!

[John Murphy]

Hi Roger,

I am very busy creating a new NSG release. This will have up to date documentation and the C++ module, Linux version, will work with Ubuntu. It also includes a bug fix that could occur with long filenames that included lots of spaces.

 

I will get back to you after this is finished. I think we can find a better solution.

I think it is likely that the bright star has not been found by the star detection, and needs a manual rejection circle adding.

That you are getting a black halo around the galaxy is not normal. I might need to look into this to check to see if it is due to your experimental technique, or some other reason.

 

Speak to you soon, John Murphy

 

[John Murphy]

15 minutes ago, Andrew Bicos said:

Hi, I just started using NSG v2.1.5 and NSGXnml but I'm having a problem that I hope is easy to solve.  However, I have spent several hours trying, so decided to ask for help.  I opened the script, add my calibrated registered images (ending in _c_cc_r from WBPP), selected a reference image that has a high altitude, entered an output directory and along with the defaults checked Gradient images, checked ImageIntegration and entered a Process icon from the ImageIntegration process using the new instance triangle onto my workspace as a placeholder and checked Auto exit and run ImageIntegration.  I kept all other defaults.  I ran this but noticed that when I opened up the Process icon on my workspace that the fields had been populated but for Normalization it had selected Local normalization, which I knew was wrong since it should be No normalization.  I then checked my output directory and saw that there were no new images with the _nsg appendage - there were only gradient images and xnml files only.  What am I doing wrong?  Thanks!

When using the NSGXnml module, NSG creates '.xnml' normalization data files. These contain all the information ImageIntegration needs to scale the images and remove the relative gradient. Hence there is no need to create corrected images (the files with the _nsg postfix).

 

ImageIntegration is populated with the registered but uncorrected images, and the '.xnml' normalization files. We now need to tell ImageIntegration to apply the instructions stored in the 'xnml' data files in order to internally create corrected images. We do this by specifying Normalization = Local normalization.

 

Applying the normalization within ImageIntegration is great because internally, ImageIntegration does not have to limit the data to the 0 to 1 range. None of the images get truncated. It is also essential if you want to use DrizzleIntegration.

 

The new Reference Documentation I am currently writing should make all this a bit clearer.

 

I really should have called NormalizeScaleGradient 'PhotometricLocalNormalization'. Probably too late to rename it now.

 

Hope this helps, John Murphy

Edited April 6, 2022 by John Murphy

[Andrew Bicos]

Yes, that helps explain a lot.  Thanks!

 

 

[sinharay]

Hi John,

    thanks for creating the latest update - it is working fine on windows.

 

    On Linux, however I am seeing a problem and I do want to use Linux since it is so much faster

 

1. i am using Kubuntu 20.0.4

2. I have PixInsight 1.8.9 and I installed it after removing the old PI installation completely

3. when i try to install the latest version - NSGXnml (Latest PixInsight 1.8.9, API version 0x173) :

•          PixInsight module install is not able to recognize the latest .so (API 0x173)
•          error message - 0 additional PixInsight module(s) were found on directory: /home/backyard_photons/PI_Install/Pixinsight_NSG (Recursive search)

 

4. the NormalizeScaleGradient.js version 2.1.5 is getting installed correctly

 

5. as an experiment I tried installing the older version of NSGXnml (PixInsight 1.8.8-12, API version 0x172) and it gets installed. 

• PixInsight is also able to load it during startup from the proper location
• i shut down PI
• i then replaced the old .so (API 0x172) with the new version of .so (API 0x173) in the same folder where PI was loading it at startup
• now when i start PI it is not able to load the .so with the following error

*** PixInsight API Error: Module: /home/backyard_photons/PI_Install/Pixinsight_NSG/NSGXnml-pxm.so
Cannot load library /home/backyard_photons/PI_Install/Pixinsight_NSG/NSGXnml-pxm.so: (/lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found

(required by /home/backyard_photons/PI_Install/Pixinsight_NSG/NSGXnml-pxm.so)): Module load error

 

 

Do I need to update the GLIBC version on my Kubuntu setup to be compatible with the NSGXnml-pxm.so?

 

regards

Shibaji

[John Murphy]

Hi Shibaji

I have just uploaded

 

PI1.8.9_NSGXnml-pxm_1.0.2_linux.tar.gz

 

to the NormalizeScaleGradient website:

 

https://sites.google.com/view/normalizescalegradient

 

This has been compiled using Kubuntu, and I believe it should work with all supported Linux versions of PixInsight.

 

Let me know how you get on.

 

The repository will be updated very soon. I am currently finishing the NSG 2.1.6 release. Within the next couple of days.

 

Regards, John

 

Edited April 7, 2022 by John Murphy

[sinharay]

Hi John,

     I installed the update and it works perfectly. Thank you so much!

 

regards,
Shibaji

[John Murphy]

Finally, I have got the Linux version sorted! It has been a big learning curve, and I spent several full days pulling my hair out!

Problems first started for me when I tried to dual boot my rather old PC for windows and Linux. I first tried Ubuntu, but it would hang during installation... I then tried Fedora, which installed without problem. So I compiled my code on Fedora. The problems occurred because Fedora is on the cutting edge. It uses the very latest versions of gcc and g++. This would then cause problems on other Linux Long Term Support variants, which often used older versions.

I finally managed to get Kubuntu to install. This initially also hung during installation, but I eventually discovered how to fix this - I had to re-partition the disk. Kubuntu uses the correct gcc and g++ versions, so the problems are now solved...

 

Download from 

https://sites.google.com/view/normalizescalegradient

 

I got there in the end!

Regards, John Murphy

Edited April 7, 2022 by John Murphy

[John Murphy]

NormalizeScaleGradient 2.1.6

Is now available on the NSG website:

https://sites.google.com/view/normalizescalegradient

The repository is currently being updated. It will probably show the update tomorrow.

 

Documentation

I have also released the Reference Documentation for NSG 2.1.6

I would recommend that everyone reads the Quick Start Guide again, because new functionality has significantly changed how the tool is used. 

 

Adam Block has just released a video tutorial of NSG 2.1.6 in his Fundamentals section. He's done a great job of explaining how to use it.

 

Recent updates include:

 

Photometry Graph

Now displays the (brightness) scale in the title bar. This is the gradient of the best fit line.

Swapped the Reference and Target axis so that the displayed gradient matches the scale.

 

Gradient dialog

This can now display the corrected target image (corrected for scale and gradient). Toggle buttons are provide to allow the user to toggle between the target and the corrected target, and between the reference and the corrected target. This provides feedback on how well the correction has worked. Adjust the gradient smoothness until the correction is sufficiently accurate. What you see is what you will get!

 

A toggle button is available to toggle between the image and graph view.

 

Increased the gradient smoothness range to -4.0 to +4.0

Changed the default gradient smoothness to -2.0

 

Reduced the size of the star rejection circles by changing the default growth rate from 1.0 to 0.25

 

Integration section

Removed the 'Output icon' combo box from the Integration section.

Added a new check box 'Reset template' (you should usually use the default of on).

The behavior of the 'Template icon' combo box has changed. If 'Reset template' is checked, the ImageIntegration template icon is only used to save the settings required by NSG. All other settings will be set to the default ImageIntegration settings. For more details, see the NSG Reference Documentation.

 

Removed the minimum weight option. This was causing confusion for some when using NSG with DrizzleIntegration. They were not always aware that some of the images were rejected due to their weight. They then added all images to DrizzleIntegration, some of which had not had their drizzle data files updated by ImageIntegration. This causes an error.

 

The process to reject images with low weights has changed. It is now far more flexible, but does require user input. That is a good thing, because it is then obvious to also omit these files from DrizzleIntegration. Once ImageIntegration is displayed, the first image is the reference. All other images are sorted by descending weight. Use the ImageIntegration tooltip to display the image weights at the end of the list. Double click on the green tick to deselect images with low weights. This makes it very easy to experiment and see how much difference removing the images is making.

 

Output section

The 'Rescale' option seemed like a good idea when I added it, but it didn't work as well as I had hoped. The resulting stacked image often had a reduced dynamic range. The highest value  could be well below 1.0 The reason for this is that if only a few images had a really good dynamic range, NSG would scale for these images. But ImageIntegration's data rejection would remove the star peak because it would be seen as a data outlier. Hence the images were scaled more than they needed to be. I am concerned that this could adversely affect the excellent MureDenoise script.

 

The 'Rescale' option also provided a false sense of security, because it does not provide any protection from the truncation of negative values.

 

Some people have found that when using NSG with a very large number of images, it can slow down after a certain number have been processed. The solution is to split the job up into batches, and restart PixInsight after each batch has completed. I don't know why this occurs, and I have not been able to reproduce it. But if a job is split into batches, and the 'Rescale' option is used, the results will not be compatible. I wanted to remove the risk of this error.

 

I have therefore removed the 'Rescale' option. The best solution to the truncation problem is to use the 'Normalization data' option. NSG then produces '.xnml' Normalization data files. These do not truncate the data. These '.xnml' data files instruct ImageIntegration how to correct both the scale and gradient. Internally, ImageIntegration is not limited to the 0.0 to 1.0 range, so no truncation occurs.

 

 

I hope you find the ability to blink between the reference and the corrected target image useful.

Regards, John Murphy

[John Murphy]

NormalizeScaleGradient 2.1.6 is now available from the NSG repository, including up to date Reference Documentation. I would recommend you read the Quick Start Guide because recommended usage has changed in this release. There is also a new section that describes how to use NSG with DrizzleIntegration.

 

A big thank you to Marco for helping me with this.

Regards, John Murphy

[John Murphy]

Hi all, I made a recent change to the ImageIntegration section:

In the previous version, I had two combo boxes: Template icon and Output icon.

 

The Output icon was used to save the ImageIntegration settings. This is useful because it adds the images in weight order, making it easier to deselect the images with the worst weights at the bottom of the list. It also sets up the correct settings - for example Local Normalization, and NWEIGHT. The Output icon's initial content is ignored, and it is always overwritten.

 

The Template icon was only used if the user wanted to specify their own initial settings, instead of changing the settings once they invoked ImageIntegration from the process icon. This process icon was never overwritten.

 

In the new version, only one icon is used. If 'Reset template' is selected, this process icon is only used for output, so it will get overwritten with ImageIntegration defaults plus the NSG settings. If 'Reset template' is not selected, the process icon will be used for input, it will be modified with NSG settings and then it will also get overwritten. Hence, if you want to keep the original process icon, you need to copy it first.

 

Should I go back to having two combo boxes? Which system is easier to understand and use? Please let me know what you think. 

Regards, John Murphy.

 

[Edit]

So far I have one vote from Adam Block to keep the new behavior, and one vote against from Linwood.

 

Perhaps I should have a once only warning before overwriting the process icon for the first time? I think once a user knows to keep a copy of their ImageIntegration icon, for example labelled NSG_ImageIntegration, the new version should work smoothly.

Edited April 13, 2022 by John Murphy

[John Murphy]

OK, I have decided to stay with the new behavior. I think most people will realize the template ImageIntegration process icon that they specified has been updated, because when they double click on it they see the updated values in ImageIntegration. It is also written to the console. I therefore think a warning dialog would be more of a nuisance than a help.

 

I would recommend that people create an ImageIntegration icon specifically for NSG. Give it a name that contains "NSG".

 

I would also recommend that all NSG users watch Adam Block's excellent new YouTube video on NormalizeScaleGradient.

Regards, John