TheSkySearchers

AstroBee wrote: ↑Mon Mar 18, 2024 2:05 am

Graeme1858 wrote: ↑Mon Mar 18, 2024 12:15 am .... I'm impressed with GraXpert. Although it crashed when I tried to put the unprocessed image in. Does it expect a non linear image?

No, it works best with a linear image.

Thanks Greg.

That's interesting, I tried the untouched, linear .fit file twice and both times the programme hung. It wasn't until I tried the the smaller file, non linear, starless image that it worked. Then it did an excellent job. Perhaps I'll try it on the linear .tif file.

Graeme

Graeme1858 wrote: ↑Mon Mar 18, 2024 12:15 am Here's mine.


Ceph_Dust - GD_01.jpg



Excellent data, except for the gradient at both ends of the image. So many stars! I tried a DBE but it just got confused due to the absence of black background. I'm impressed with GraXpert. Although it crashed when I tried to put the unprocessed image in. Does it expect a non linear image?

Good fun. Thanks for the data David.

Graeme

Thats looking good, you have all the main dust showing and the red Ha too. I'd probably run a SCNR on it as it has a bit of a green cast, seen especially around the bright star at lower middle.

I dont know why GraXpert crashed on you. As @AstroBee Greg says, its best to do it as ealy as possible. I do it as the very first step (unless a crop of edges is required. Are you using it in AI mode - thats the only way I've ever used it. Otherwise I dont see any benefit of it over ABE/DBE. Actually, For this image ABE did a better job than DBE (not as good as GX). DBE took a lot of work comparing to images on Astrobin to find background areas etc.

Heres a before and after GraXpert of the shared fits so you can check your settings. It really does an incredible job



BTW, its possible to integrate GX into Pixisight using a script so all you have to do it open the scropt (toolbox-Graxpert) and off you go. Its handy as you dont have to export and re-import the image. You can get the repository and details of setting it up from Lukomatico here

Cheers David.

I'll be integrating that then. And I shall be watching a few videos on the laptop today as well.

Graeme

This is my attempt. I'm still new to Pixinsight so I can't figure out the pixelmath to bring the starless nebula in front of the stars. This is some tough data for sure.

This is my Processing:
Dynamic Crop
Graxpert (I love this script)
ColorCalibration (I couldn't do SPCC because there was no solve information)
BlurXterminator
NoisXterminator
Histogram Transformation
StarXterminator
A) starless - Curves
SCNR
CorrectMagentaStars
B) stars - Curves
SCNR
CorrectMagentaStars
PixelMath - "starless+stars"

This is my Starless version of the image.

That's a fine process Ron. You have brought out the dust really well. Did the Color Calibration Process have a good effect?

I normally use Photometric Color Calibration but I forgot to add it to my workflow this time! Think I'll have another go!

Graeme

BigKahuna wrote: ↑Tue Mar 19, 2024 2:08 am This is my attempt. I'm still new to Pixinsight so I can't figure out the pixelmath to bring the starless nebula in front of the stars. This is some tough data for sure.

This is my Processing:
Dynamic Crop
Graxpert (I love this script)
ColorCalibration (I couldn't do SPCC because there was no solve information)
BlurXterminator
NoisXterminator
Histogram Transformation
StarXterminator
A) starless - Curves
SCNR
CorrectMagentaStars
B) stars - Curves
SCNR
CorrectMagentaStars
PixelMath - "starless+stars"

SharkNebulaAA.jpg

Thats really good Ron, I particularly like how you have the dust around the rotting fish looking so grey and "misty".

On image solving - as this is a fits file the information needed to solve the image is embedded in the fits header - so simply running the ImageSolver script would open it up already populated with the necessary information and you just need to hit execute.

However, to @Graeme1858 's point, I really dont know how much difference CC (in any form) makes to this type of image. Take my original PI image (SPCC) compared to my Photoshop (No form of CC) image posted above as an example. There is very little to draw between them in terms of colours and what difference there is is probably from my processing rather than as a result of CC. Its one of these things I do because it seems to be expected, rather than from any strong belief that it makes much difference

For star removal and replacement you should take a look at Bill Blanshans "Screenstars" script - I'll link below.. Using that you remove the stars from one copy (using starxterminator) then its takes your still starry copy and the new starless copy, does some clever, and frankly beyond level of understanding stuff, and spits out a stars only image. Once you have done all your further processing you go back to the screen stars script and it again does some clever stuff and gives you your final image with stars. I find it really does make a difference compared to trying to do it via pixelmath - the stars look significantly better using the script.

Jockinireland wrote: ↑Tue Mar 19, 2024 9:26 am For star removal and replacement you should take a look at Bill Blanshans "Screenstars" script

Bill Blanshan and Mike Cranfiled have done some excellent Scripts. I use the Star Reduction Script on my images.

Graeme

Well, thank you both for the comments. I've only been using Pixinsight for a little over a month and it is a "work in progress". Lol. That is why this is such a great hobby. This was a great learning experience and I appreciate you giving up your data for this challenge

I usually use imagesolver for my images because I use DSS to stack instead of Pixinsight. I should have tried it but color calibration did a "good enough" job.

I found that a very difficult image to process, particularly bringing out the dust. Could not do anywhere near as good as Jock.

Processed in Photoshop.

Hi Carole,

I'm glad you took a shot at it and I agree, its not an easy one to process - the density of stars and the minimal background dont help. But I think you've done pretty good with it. You have great colour in the stars.

Can I ask, are you using any of the additional tools outside of photoshop, like GraXpert and starnet ++. If not, how did you do gradient removal on this because you've cleaned my horrible gradient very well. And do you have a star removal step in your photoshop process?

BigKahuna wrote: ↑Tue Mar 19, 2024 10:31 am Well, thank you both for the comments. I've only been using Pixinsight for a little over a month and it is a "work in progress". Lol. That is why this is such a great hobby. This was a great learning experience and I appreciate you giving up your data for this challenge

I usually use imagesolver for my images because I use DSS to stack instead of Pixinsight. I should have tried it but color calibration did a "good enough" job.

I think at a month into pixinsight you are doing great. I dont thinks its as a scary and difficult as its often made out to be, is it?

Can I ask why you dont stack in Pixinsight? I simply never got on with DSS and moved early on to APP which I found did a brilliant job and was easier to use than DSS. When I first got PI I continued to use APP for several months then I bit the bullet and tried WBPP in Pixinsight. I found that once i understood how to set it up, which was not that hard, it was an absolute breeze to use and worked just as well as APP. Now it was a bit slower but i think thats because its doing some stuff that APP didnt do.

Can I ask, are you using any of the additional tools outside of photoshop, like GraXpert and starnet ++. If not, how did you do gradient removal on this because you've cleaned my horrible gradient very well. And do you have a star removal step in your photoshop process?

I don't have anything outside of Photoshop and I do a tiny amount of star size reduction in Images plus.
Gradient removal is Gradient Exterminator in PS. I have Annie Actions which I use for Star removal within photoshop, but I have a lot of photoshop tricks up my sleeve from many years picking up tips here and there. One in particular I learned from Ollie Penrice, which is Exploiting the Equalize Function in Photoshop. It is the only routine where I use Masks. I have another routine for reduction of star sizes in photoshop (but it does leave a lot of minute stars behind so have other tricks up my sleeve for that too. Finally a routine for noise reduction learnt off Roberto Colombari on Astrobin.

HTH

Carole

Jockinireland wrote: ↑Tue Mar 19, 2024 11:19 am
I think at a month into pixinsight you are doing great. I dont thinks its as a scary and difficult as its often made out to be, is it?

Can I ask why you dont stack in Pixinsight? I simply never got on with DSS and moved early on to APP which I found did a brilliant job and was easier to use than DSS. When I first got PI I continued to use APP for several months then I bit the bullet and tried WBPP in Pixinsight. I found that once i understood how to set it up, which was not that hard, it was an absolute breeze to use and worked just as well as APP. Now it was a bit slower but i think thats because its doing some stuff that APP didnt do.

I am told that time to stack in pixinsight is MUCH longer than outside software. Plus, I've been using DSS for 5 years now so familiarity with it is another reason.

carastro wrote: ↑Tue Mar 19, 2024 12:34 pm

Can I ask, are you using any of the additional tools outside of photoshop, like GraXpert and starnet ++. If not, how did you do gradient removal on this because you've cleaned my horrible gradient very well. And do you have a star removal step in your photoshop process?

I don't have anything outside of Photoshop and I do a tiny amount of star size reduction in Images plus.
Gradient removal is Gradient Exterminator in PS. I have Annie Actions which I use for Star removal within photoshop, but I have a lot of photoshop tricks up my sleeve from many years picking up tips here and there. One in particular I learned from Ollie Penrice, which is Exploiting the Equalize Function in Photoshop. It is the only routine where I use Masks. I have another routine for reduction of star sizes in photoshop (but it does leave a lot of minute stars behind so have other tricks up my sleeve for that too. Finally a routine for noise reduction learnt off Roberto Colombari on Astrobin.

HTH

Carole

Thanks Carole. When I did my one in photoshop I looked at using Gradient Exterminator - I got it some years ago when I used PS more - but could not work out how to use it with so little actual background. So I put the fits through GraXpert instead.

Hmmm you have got me thinking as l just did GE on the entire image almost at the beginning.

Maybe that is why l found it so hard to bring out the dust. Maybe l will have another go at it. Might do it in layers and only erase parts of it.

Thanks for sharing this, David. I loved this image when I saw it on AB.

Here's my quick take. I'm not sure whether the circle like smudge is a DBE artifact or maybe a dust donut. Not quite sure how I'd process it out - any ideas anyone? I probably dialed back the stars too much here, but there are so many! It would be worth stretching just the bright ones up a bit.
Processed in PI
BlurX - Color Only
DBE (tried new gradient correct tool and failed miserably)
Blurx - Sharpen non stellar
Clone version for Stars
StarX
GHS - multiple passes
Stretch stars copy in HT
Remove stars and then screen back in using Mike Cranfields Screen stars script
Color tweak in curves.

starfield wrote: ↑Tue Mar 19, 2024 6:12 pm Thanks for sharing this, David. I loved this image when I saw it on AB.

Here's my quick take. I'm not sure whether the circle like smudge is a DBE artifact or maybe a dust donut. Not quite sure how I'd process it out - any ideas anyone? I probably dialed back the stars too much here, but there are so many! It would be worth stretching just the bright ones up a bit.

MarchAPChallenge3@0.5x.jpg

Processed in PI
BlurX - Color Only
DBE (tried new gradient correct tool and failed miserably)
Blurx - Sharpen non stellar
Clone version for Stars
StarX
GHS - multiple passes
Stretch stars copy in HT
Remove stars and then screen back in using Mike Cranfields Screen stars script
Color tweak in curves.

I think thats really good.

I was hoping someone would try the new gradient correct on it. I've not updated PI yet for it because I had so many problems with the last three builds. So interested to hear it did not handle this well. And to be honest I find GraXpert does a great job anyway.

I'm not sure where your dust bunny is - could you point it out so I can investigate.

thanks

David.

Jockinireland wrote: ↑Tue Mar 19, 2024 7:59 pm

starfield wrote: ↑Tue Mar 19, 2024 6:12 pm Thanks for sharing this, David. I loved this image when I saw it on AB.

Here's my quick take. I'm not sure whether the circle like smudge is a DBE artifact or maybe a dust donut. Not quite sure how I'd process it out - any ideas anyone? I probably dialed back the stars too much here, but there are so many! It would be worth stretching just the bright ones up a bit.

MarchAPChallenge3@0.5x.jpg

Processed in PI
BlurX - Color Only
DBE (tried new gradient correct tool and failed miserably)
Blurx - Sharpen non stellar
Clone version for Stars
StarX
GHS - multiple passes
Stretch stars copy in HT
Remove stars and then screen back in using Mike Cranfields Screen stars script
Color tweak in curves.

I think thats really good.

I was hoping someone would try the new gradient correct on it. I've not updated PI yet for it because I had so many problems with the last three builds. So interested to hear it did not handle this well. And to be honest I find GraXpert does a great job anyway.

I'm not sure where your dust bunny is - could you point it out so I can investigate.

thanks

David.

I may be imagining it.... but this is where I sort of see it.....

Hi all

Thanks for sharing the data I really enjoyed processing this. My steps are as follows

Pixinsight
Image solver for SPCC
Gradient Correction (new PI process)
BlurXterminator
NoiseXterminator
StarXterminator
GHS to stretch the starless version

Photoshop / Camera Raw
Worked on the starless image using colour sliders to get the nebulosity the colour I wanted
Some colour noise reduction

Pixinsight
On the stars image I used histogram transformation to stretch the starts to the level I wanted
PixelMath to rescreen the stars
Selective colour correction script again to obtain colours and look I wanted
colour mask to try and saturate the Ha below VDB152
Final curves adjustment, saturation & contrast

I also decided to go against the norm with the mirror flipped version.

Thanks Simon