Horsehead old vs new

[Rodd]

I was never satisfied with this image due to artifacts born from over sharpening. I managed to sharpen the breath of the horse, but caused issues elsewhere.  Then I read the Deconvolution tutorial and decided to give it a whirl.  I think the image is much better for it.  Alnitak is sounder (no stippling in the glare).  Conditions were not great so the star is not as tight as it could be--but it still split. The circular blue reflection artifact at right is less pronounced (Probably because I combined the Ha differently....correctly this time).   NGC 2023 is just as defined--though I like the color in the old image….but I think the palette in the new image is more correct.  Also, the color of the Flame Nebula is closer to what it should be.  

But finally, the details of the image are sharpened even more without the artifacts.  The horses breath is more distinct.  I know there are very faint dark rings around some stars--only visible when you zoom and really look.  They are the result of not using a strong enough dark deranging setting in Deconvolution.  Fixing that means undoing quite a bit of work and redoing deconvolution of the synthetic luminance, then reinserting and redoing a bunch of stuff.  If that is really necessary, I will do it (I am hoping the consensus will be it isn't)--but don't let that stop you from being critical.  I know there is room for improvement (like the faint greenish cast in places--SCNR was used). Please let me know if you think I am bonkers.

Image 1:  The original

Image 2: The new version using Deconvolution in Linear state.

[Rodd]

Fixed the greenish cast.   No more tinkering ( and a tad sharper!)

 

 

[Sunshine]

A beautiful image it is!

[Rodd]

7 minutes ago, Sunshine said:

A beautiful image it is!

Thanks Sunshine.    Long in coming!

Rodd

[Tim]

Better Rodd, although the original was still very nice. Your reprocessing has brought more life and depth to the foreground region underneath the horsey, the region that in fact gives birth to the feature, so nice to make something of it instead of just a muddy area.

It's one area I come back to year on year and spend many hours on, always tinkering too. One of these times we might be happy to move on

In fact, I still haven't processed last winters efforts, really should get round to that!

Cheers

Tim

[Rodd]

30 minutes ago, Tim said:

Better Rodd, although the original was still very nice. Your reprocessing has brought more life and depth to the foreground region underneath the horsey, the region that in fact gives birth to the feature, so nice to make something of it instead of just a muddy area.

It's one area I come back to year on year and spend many hours on, always tinkering too. One of these times we might be happy to move on

In fact, I still haven't processed last winters efforts, really should get round to that!

Cheers

Tim

Thanks Tim..  I agree, there is always something to work on.   

Rodd

[celestron8g8]

I agree , better and nice capture !

[Rodd]

1 minute ago, celestron8g8 said:

I agree , better and nice capture !

Thanks Ron.   Yay!

Rodd

[wimvb]

Great rework. Nothing to be critical about, imo.

2 hours ago, Rodd said:

Fixing that means undoing quite a bit of work and redoing deconvolution of the synthetic luminance, then reinserting and redoing a bunch of stuff.  If that is really necessary, I will do it (I am hoping the consensus will be it isn't)--but don't let that stop you from being critical.

If you did all the processing in PixInsight, redoing an image is very easy. Just use the process history. All the settings are saved, together with which mask you used. But I guess you knew that.

[Rodd]

17 minutes ago, wimvb said:

Great rework. Nothing to be critical about, imo.

If you did all the processing in PixInsight, redoing an image is very easy. Just use the process history. All the settings are saved, together with which mask you used. But I guess you knew that.

Thanks Wim.  I don't usually save my work as projects.  Don't know why--I guess it just how I started.  It does take up allot more disc space.,

Rodd

[wimvb]

Yes, it takes up a lot of space, but I find it worth it. Otoh, with the large number of projects you have (done recently), I can understand the space concern.

The last few years, we've had about 4 imaging nights per month, on average. And from mid April until mid August, imaging isn't possible anyway. This means I don't have that many projects to save. ?

[Rodd]

23 minutes ago, wimvb said:

Yes, it takes up a lot of space, but I find it worth it. Otoh, with the large number of projects you have (done recently), I can understand the space concern.

The last few years, we've had about 4 imaging nights per month, on average. And from mid April until mid August, imaging isn't possible anyway. This means I don't have that many projects to save. ?

Quality is better than quantity anyway....and you have that for sure.  But I feel for you.  When it is cloudy for weeks I get frustrated.  I hope things return to how they were a couple years ago.

Rodd

[wimvb]

6 hours ago, Rodd said:

Quality is better than quantity anyway...

Hehe, in AP quantity IS quality. (But not the way you meant it.) Soon darkness will be back here, and during winter we have long nights to compensate.

In the mean time, just keep those pretty pictures coming. I enjoy them.

[newbie alert]

At first I thought you was bonkers..only a slight change, but the more I look the more I see...id still be ecstatic to achieve the first image...

[alexbb]

Rodd, that looks excellent!

12 hours ago, Rodd said:

Fixing that means undoing quite a bit of work and redoing deconvolution of the synthetic luminance, then reinserting and redoing a bunch of stuff.  If that is really necessary, I will do it (I am hoping the consensus will be it isn't)

I don't master this yet, but you can save all your actions in a process. You can deconvolve again the image, then apply the process on the result.

[Rodd]

3 hours ago, newbie alert said:

At first I thought you was bonkers..only a slight change, but the more I look the more I see...id still be ecstatic to achieve the first image...

Thanks Newbie!  Yeah, not huge changes--but its amazing how tiny incremenets can be significant with AP images.

Rodd

[Rodd]

1 hour ago, moise212 said:

Rodd, that looks excellent!

I don't master this yet, but you can save all your actions in a process. You can deconvolve again the image, then apply the process on the result.

Thanks Alex,  I do have (did have) the deconvolution saved as an icon that I could then just drag to the image and it would do it--but that action has to be done in the linear state.  So if I were to do more deconvolution now, I would have to go back (undo) everything after the stretch to apply a second iteration of Decon.  Then I would have to redo everything.  I don't save my work in projects--which I probably should do--so I would have to look at the history and write down all the steps--remembering the settings as they are not saved in the history explorer pull down table.  I don't have allot of room on my hard drive due to the CMOS style of imaging (lots of subs), and saving as projects eats up allot of space.

Rodd

[alexbb]

26 minutes ago, Rodd said:

Thanks Alex,  I do have (did have) the deconvolution saved as an icon that I could then just drag to the image and it would do it--but that action has to be done in the linear state.  So if I were to do more deconvolution now, I would have to go back (undo) everything after the stretch to apply a second iteration of Decon.  Then I would have to redo everything.  I don't save my work in projects--which I probably should do--so I would have to look at the history and write down all the steps--remembering the settings as they are not saved in the history explorer pull down table.  I don't have allot of room on my hard drive due to the CMOS style of imaging (lots of subs), and saving as projects eats up allot of space.

Rodd

I believe that's exactly what the process container does. You drag an action to the process container and it will be stored with the parameters. So you deconcolve the image in the linear phase, you then do all the processing, but additionally, you drag each applied action to the process container. Then, if you want to start over, you only deconvolve again the linear image and then you apply the processes in the container at once. You'd need perhaps to have the masks.

[Rodd]

3 minutes ago, moise212 said:

I believe that's exactly what the process container does. You drag an action to the process container and it will be stored with the parameters. So you deconcolve the image in the linear phase, you then do all the processing, but additionally, you drag each applied action to the process container. Then, if you want to start over, you only deconvolve again the linear image and then you apply the processes in the container at once. You'd need perhaps to have the masks.

Ahh... Thanks--I did not know this.  I usually have all the masks still open while I am processing so that would work.  Thank you...I will start exploring the process container.

Rodd