Sh2 -101 Widefield

[carastro]

Quote

just to be clear, it was the Morphology causing the artifacts not the noise reduction, in case anyone else was wondering. 

I have no idea what MT is, I don't use Pixinsight, but whatever it is seems to have had the same effect as OTT noise reduction.  But anyway the image certainly looks better without using MT.

Carole 

Edited July 24, 2019 by carastro

[steviemac500]

@carastro, sorry I wasn’t being offensive, I was trying to be funny about how this post kept going round the same conversation.

Morphological Transformation is usually used to dilate or erode stars. It can create a splotchy effect with the filaments present if it’s used badly, as in my image. I tend to use it to draw more attention to the nebula but I may not anymore. 😀

[Datalord]

PI is incredibly effective at so many tasks. The learning curve is to figure out which process to use when and, more important, which mask to create for what purpose. 80% of success comes down to masks, but once you get a grip on that, you can do some crazy things with it.

[Xplode]

I use mostly PI too.
One thing i've found is that if doing a process and it's effect is too drastic and it's not possible to tune it down it can be a nice thing to combine the before and after image with Pixelmath to give a more pleasing result, it does of course not work with all processes.

[glowingturnip]

23 hours ago, Xplode said:

I use mostly PI too.
One thing i've found is that if doing a process and it's effect is too drastic and it's not possible to tune it down it can be a nice thing to combine the before and after image with Pixelmath to give a more pleasing result, it does of course not work with all processes.

indeed, almost everything I do in PI I merge back with a 'before' version, experimenting with the opacity to suit.  Saved process icon for it, but would be nice if they actually formalised a tool, real time sliders etc - but then it gets too much like PS for their tastes

Edited July 25, 2019 by glowingturnip

[Xiga]

On ‎24‎/‎07‎/‎2019 at 12:45, steviemac500 said:

@carastro, sorry I wasn’t being offensive, I was trying to be funny about how this post kept going round the same conversation.

Morphological Transformation is usually used to dilate or erode stars. It can create a splotchy effect with the filaments present if it’s used badly, as in my image. I tend to use it to draw more attention to the nebula but I may not anymore. 😀

I'm not a PI user myself. It just looks unnecessarily convoluted to my eyes, and don't get me started on all those names lol 🙄 

But just on the subject of star reduction, rather than use MT, you might want to look into a new technique that a certain Adam Block came up with:

https://pixinsight.com/forum/index.php?topic=13567.0

Edited July 26, 2019 by Xiga

[wimvb]

On 23/07/2019 at 19:50, Datalord said:

The alternative is to make a completely starless version, but I don't know how to help you with that.

Starnet++ is now a module in PI. That should do the trick. No sliders, no mask just select and apply.

[wimvb]

On 23/07/2019 at 19:50, Datalord said:

 

 Sorry, double post.

Edited July 26, 2019 by wimvb

[Datalord]

1 minute ago, wimvb said:

Starnet++ is now a module in PI. That should do the trick. No sliders, no mask just select and apply.

How do you get that?

[wimvb]

Have a look here

https://pixinsight.com/forum/index.php?topic=13691.0

It's available from sourceforge.

[wimvb]

On 24/07/2019 at 11:22, ollypenrice said:

Still, if you like trying to mend your watch while wearing boxing gloves you'll love PI!

lly

That's exactly what PS feels like to me. 😄

Edited July 26, 2019 by wimvb

[Datalord]

18 minutes ago, wimvb said:

Have a look here

https://pixinsight.com/forum/index.php?topic=13691.0

It's available from sourceforge.

Wow, I think I saw his work on astrobin earlier. I was wondering if he would make a PI implementation. This is insanely cool.

[wimvb]

@steviemac500, I just had a play with the image you posted. I created a star reduced and a starless version (starnet++)

I think that your star reduction didn't work because of the mask. I created a ring mask by combining two ordinary star masks:

A. Noise threshold = 0.5, Layers = 6, Large scale/Small scale/Compensation = 2/1/2, Smoothness = 10, Aggregate and Binarize checked

B. Same settings but Large scale/Small scale/Compensation = 0/1/0 and smoothness = 3. Strengthened this mask by bringing in the white point somewhat.

Then pixelmath to create the ring mask: A - B

Morphological Transform: 3x3 element, Morph. Selection, amount = 0.5, selection = 0.25, iterations = 8

The starless version was created with standard settings in starnet++ on the original image.

Combined with PixelMath, applied to the star reduced image:

iif((X()+Y())>1, $T, starless)

 

[Datalord]

34 minutes ago, wimvb said:

The starless version was created with standard settings in starnet++ on the original image.

OMFG! WITCHCRAFT!

[wimvb]

11 minutes ago, Datalord said:

OMFG! WITCHCRAFT!

No, just Artificial Intelligence at work 🤓

[steviemac500]

6 hours ago, wimvb said:

@steviemac500, I just had a play with the image you posted. I created a star reduced and a starless version (starnet++)

I think that your star reduction didn't work because of the mask. I created a ring mask by combining two ordinary star masks:

A. Noise threshold = 0.5, Layers = 6, Large scale/Small scale/Compensation = 2/1/2, Smoothness = 10, Aggregate and Binarize checked

B. Same settings but Large scale/Small scale/Compensation = 0/1/0 and smoothness = 3. Strengthened this mask by bringing in the white point somewhat.

Then pixelmath to create the ring mask: A - B

Morphological Transform: 3x3 element, Morph. Selection, amount = 0.5, selection = 0.25, iterations = 8

The starless version was created with standard settings in starnet++ on the original image.

Combined with PixelMath, applied to the star reduced image:

iif((X()+Y())>1, $T, starless)

 

Thanks Wim, I’ll have a look at the other file later. Your mask generation is very interesting, I don’t understand it, but it’s interesting. Why the 3x3 element, everything I’ve read has said to use a 5x5? 

Funny, I like starless nebula images but for some reason in Cyprus, no one else does! 🙄

By the way, what language is that bottom line written in and where do you go to learn it?

[wimvb]

2 hours ago, steviemac500 said:

By the way, what language is that bottom line written in and where do you go to learn it?

Pixelmath. Check this site:

http://www.werbeagentur.org/oldwexi/PixInsight/PixInsight.html

[steviemac500]

Thanks for all the support. The latest edition of Astronomy Now has published the image. Well chuffed!!

[Stub Mandrel]

I'm afraid there's still something really wrong with that image (I hesitate as I have never produced anything in this league).

This is a crop from centre bottom, the fault should be self-evident:

 

Basically - two panes haven't been properly aligned.

[Gina]

I used  Photoshop until the computer it was on died.  I bought the PC with PS already installed.  Now with a newer PC with much better performance I have PixInsight which I bought.  Very powerful but one hell of a learning curve.  Mind you PS was almost as bad!  I think PI is the more powerful and does everything.

[steviemac500]

11 hours ago, Stub Mandrel said:

I'm afraid there's still something really wrong with that image (I hesitate as I have never produced anything in this league).

This is a crop from centre bottom, the fault should be self-evident:

 

Basically - two panes haven't been properly aligned.

Yep, I see it now, can't believe I missed it but in fairness you have to look pretty hard or be looking for something anyway to see that!! I never noticed it first time round and I don't think it detracts from the image if you're not looking too hard. At a loss to explain it though as APP normally is faultless. Don't ever hesitate to offer criticism though otherwise how do we learn and this will be something I will scrutinise more in the future,

Still, AN have published it so I am going to remain happy with that!

Edited September 27, 2019 by steviemac500

[Stub Mandrel]

I hadn't seen the second page when I posted that or I might have kept schtumm.

I zoomed in trying to see the 'filaments' and it struck me immediately, I'm still not entirely sure what the filaments were....