Process ordering for Blur Xterminator, Noise Xterminator and SPCC

[edarter]

Hi,

While evaluating BXT & NXT to see if I want to splash the cash I've been going over old data to try them. One such dataset is Andromeda. While doing so though I have found a bit of a funny that has me wondering if there is a prescribed order in which I should be running the above processes? The reason I ask is that I was running BXT & NXT before SPCC but this resulted in an image that was VERY red biased and the SPCC plots were all over the place. After some trial and error it would seem that NXT is the culprit, needing to be run after SPCC in order to keep a well balanced image. Is this to be expected or am I doing something really dumb here? 
Example screenshots below

Actually - in uploading the screenshots I've noticed its even more subtle in that its only if I run NXT immediately before SPCC?!? Though the image where I did SPCC-BXT-NXT looks more washed out than BXT-SPCC or BXT-SPCC-NXT, but tis still way better than BXT-NXT-SPCC which was a red mess!

BXT-NXT-SPCC:

 

SPCC-BXT-NXT:

 

BXT-SPCC:

 

BXT-SPCC-NXT:

Edited January 6, 2023 by edarter
Titles of images

[scotty38]

SPCC should be done before BXT. Russ Croman advises this too.

[tomato]

I currently run SPCC at the start, then BXT, NXT is not used until the end of processing.

[Budgie1]

54 minutes ago, tomato said:

I currently run SPCC at the start, then BXT, NXT is not used until the end of processing.

+1 for the above approach.

I run BTX right before the stretch and NTX after everything else, so that it takes care of any noise introduced by processes after the stretch.

[Lee_P]

1 hour ago, tomato said:

I currently run SPCC at the start, then BXT, NXT is not used until the end of processing.

 

8 minutes ago, Budgie1 said:

+1 for the above approach.

I run BTX right before the stretch and NTX after everything else, so that it takes care of any noise introduced by processes after the stretch.

Ooh interesting, I generally run a bit of NXT fairly early on. My logic is to try and squash the noise before stretching. Stretching noise is bad, right? But you two only run NXT at the very end? Maybe I need to give that approach a go.

[Budgie1]

1 hour ago, Lee_P said:

 

Ooh interesting, I generally run a bit of NXT fairly early on. My logic is to try and squash the noise before stretching. Stretching noise is bad, right? But you two only run NXT at the very end? Maybe I need to give that approach a go.

I used to de-noise right before the stretch but I found that the later processes introduced a little more noise into the image, so it really needed another de-noise later on. So, before BXT came out, I started leaving the de-noise process until last and that removed all the noise I wanted. NXT just makes the process easier.

Edited January 7, 2023 by Budgie1

[edarter]

Interesting, so would there be scope to run nxt before the stretch then a small amount later on towards the end? I too thought killing noise early was a good thing but have noticed subsequent processes introducing noise back in...

[tomato]

Yes, on the basis you only want to run noise reduction once, I do it at the end after all of the noise has been generated.

[geeklee]

20 hours ago, scotty38 said:

SPCC should be done before BXT. Russ Croman advises this too.

This is interesting - is that in the documentation or noted online?

Given that star colours seem to change when running BlurXterminator (a patch is being worked on I'm sure I read on AB/CN) I've been running BlurXTerminator after DynamicCrop and then running SPCC.  I'll need to try the above 👍

[scotty38]

4 hours ago, geeklee said:

This is interesting - is that in the documentation or noted online?

Given that star colours seem to change when running BlurXterminator (a patch is being worked on I'm sure I read on AB/CN) I've been running BlurXTerminator after DynamicCrop and then running SPCC.  I'll need to try the above 👍

I've seen more recent comments but need to find them again. In the meantime here's one link, scroll down for the comment from Russ specifically:

https://pixinsight.com/forum/index.php?threads/spcc-after-blurxterminator.19928/#post-121433

[geeklee]

4 minutes ago, scotty38 said:

In the meantime here's one link, scroll down for the comment from Russ specifically:

Thanks.  I see he mentions that unintended colour shift on some data too - and working on a fix in AI2.

[scotty38]

Yes there is a thread on AI2 but his latest comment re order of use doesn't seem to be in that one although I think he referenced the thread I shared anyway.

[scotty38]

@geeklee  The AI2 thread I mentioned and you can get the Beta from here too if you wish:

https://pixinsight.com/forum/index.php?threads/blurxterminator-ai2-beta.20019/#post-121769

[geeklee]

1 minute ago, scotty38 said:

@geeklee  The AI2 thread I mentioned and you can get the Beta from here too if you wish:

https://pixinsight.com/forum/index.php?threads/blurxterminator-ai2-beta.20019/#post-121769

Nice one, thanks @scotty38 😊