Why so many square stars & dustless dust!

[Sp@ce_d]

First time I've tried this target but this doesn't look right to me.

Got over an hours worth of data for each LRGB channel so far.

Just doing a quick process on the RGB data in PixInsight to see what its like. I've never noticed so many square stars showing up before. Also, doesn't seem to be the dust detail in there I was expecting but maybe I need more data & a darker site? Tried DBE tool in PI but so far its just pulling out a flat pink subtraction & leaving a green target image. First time it's not worked for me, odd.. think I'll have to revisit Harry's Tutorial! 

Anyway, any ideas how best to deal with the square stars (& maybe the dust)? Otherwise I'm probably wasting time collecting more data on this target unless I do a close crop.

RGB 14 x 300 each channel

Esprit 80, 314L+, Baader filters. @ 3.32 Arcsecs/Pix

The PI process trail is in the filename...

[JohnC64]

I guess the first port of call would be to have a look at the individual subs. - Is it ok in R, G & B or is one of them out?

And then take a look in the R, G & B Stacks.  That will at least show if it is some individual dataset.

If it occurs in all the  initial subs then I am at a loss. 

John

[Sp@ce_d]

Yeah, to me it looks worse in the B integration subs. I'm just checking the ones from the other shooter (ZS71) that was running at the same time as I haven't looked through those yet.

Here's the FITS for the combined & the registered .fits for the RGB. Plus the unregistered L that I've not played with yet.

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/RGB_integration_Cp.fit

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/R_integration_r.fit

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/G_integration_r.fit

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/B_integration_r.fit

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/L_integration.fit

[Sp@ce_d]

I'm pretty sure the individual subs look ok. Something weird in the stacking I reckon. It's there in the RGB integration before any processing. I don't do calibration any more.. Even bias seemed to add noise, so not that. I'll try a stack in something else.

[JohnC64]

Its good that the individual subs are OK  and looking in the corners, the spacing looks good for the flattener too :-)
Good look with the reprocess
John

[swag72]

Regarding square stars - with some combinations they can be a fact of life. I hope someone comes in soon and says a little more, something about focal length and pixel size I think.

[JohnC64]

Theres a nice article here on undersampling. But The Esprit and the 314L  should be a good match.

http://www.astro-imaging.com/Tutorial/MatchingCCD.html  

Guess the seeing could scupper things though

John
 

[Sp@ce_d]

Yes I was wondering if it's the sample rate.. being 3.32 Arcsecs/Pix but I've not noticed it before on so many stars.

I've just run the subs through DSS (not used it for ages) Unfortunately DSS seems to crash when I try & align the stacked RGB channels though. So, anyway I've run the DSS stacked RGB through StarAlignment in PI & the .r files are starting to show bits missing on some stars. I think its possibly happening at this point.. I'll pop them onto dropbox in a minute..

[JohnC64]

You could always try the stacking on Nebulosity and see what it makes of the stack. Lots of clicking on stars though 14 x 4
Its a weird one

[Sp@ce_d]

Ok. check the stars in the second one regged in PI against the DSS stack

These files are 16MB in size BTW..

DSS B Stack

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/B_Stack.FTS

PI regged B stack

https://dl.dropboxusercontent.com/u/32747771/Square_Stars/B_Stack_r.fit

[Sp@ce_d]

You could always try the stacking on Nebulosity and see what it makes of the stack. Lots of clicking on stars though 14 x 4

Its a weird one

Yes.. why I have that as last on my list to try..

[LukeSkywatcher]

I'm purely an observing astronomer. Looking at that image............i'm thinking to myself "what do you mean by square stars and getting rid of the dust".

I see no square stars and the dust is part of the object. 

Its a great image. Leave it alone and be very proud of it.

[JohnC64]

a quick and dirty stretch and a zoom in appear to confirm that the PI stack is more "squarey".  This is very odd. I can only suggest reregistering them again and then stacking them.  
 

[Sp@ce_d]

I'm purely an observing astronomer. Looking at that image............i'm thinking to myself "what do you mean by square stars and getting rid of the dust".

I see no square stars and the dust is part of the object. 

Its a great image. Leave it alone and be very proud of it.

He he.. cheers Paul, I mean I want the dust to look "dusty" & not just a dark smudge.. I know what you mean though but I'm a fussy so & so..

[JohnC64]

and the dust is there  just hiding in between the pixels - Well if my rough and ready murdering of your image is right :-)

 

[Sp@ce_d]

and the dust is there  just hiding in between the pixels - Well if my rough and ready murdering of your image is right :-)

Sp@cedIris.png

Ah.. It lives... my beloved dust...

Well, running the stacks through RegiStar against the R still makes the B & G regged version stars look blocky. I'm going to call it a night on this for now as me heed is hurting.

[LukeSkywatcher]

He he.. cheers Paul, I mean I want the dust to look "dusty" & not just a dark smudge.. I know what you mean though but I'm a fussy so & so.. 

 

Try all you want, but some nebulae are Dark Dust and there is nothing you can do to change that.

 

If anything.............i'd be adding exposures to the image to enhance the dark dust.

[Naemeth]

Could it be a thin layer of mist which makes the brighter stars appear to stretch out slightly?

Lovely image though, square stars or not

[Mike Hawtin]

I've had this problem with "square" stars in the past and the fix for me was to change the default settings when running Image Registration.  Go the the Interpolation dropdown at the bottom of the  Star Alignment tool and change the Pixel interpolation to Bicubic spline and reduce the clamping threshold value, in my case I had to reduce it to 0.10 before I got round stars in the stack.

[Sp@ce_d]

I've had this problem with "square" stars in the past and the fix for me was to change the default settings when running Image Registration.  Go the the Interpolation dropdown at the bottom of the  Star Alignment tool and change the Pixel interpolation to Bicubic spline and reduce the clamping threshold value, in my case I had to reduce it to 0.10 before I got round stars in the stack.

Mike, I take my hat off to you sir.

That's sorted it

Been using PI for a year now and it's the first time I've noticed this problem. Must be the new Esprit / 314L+ combo sample rate highlighting it.

Thanks for that.. appreciated

Graham

Before...

After...

[Mike Hawtin]

No problem, happy to help but I can't take credit for the fix, I had to ask the same question on the Pixinsight forum a couple of months ago.

[ollypenrice]

The image has a big green gradient, worse on the upper left. I might try running SCNR green before DBE, but try it both ways. In DBE try a model with not many sample points. It should just pick up gradients with no visible structure, just a drift in colour.

This really is faint dust and prone to picking up gradients from the background sky, plastics factories and Chinese lanterns, whatever!!!

Olly

[Sp@ce_d]

The image has a big green gradient, worse on the upper left. I might try running SCNR green before DBE, but try it both ways. In DBE try a model with not many sample points. It should just pick up gradients with no visible structure, just a drift in colour.

This really is faint dust and prone to picking up gradients from the background sky, plastics factories and Chinese lanterns, whatever!!!

Olly

Cheers Olly, I'll try that. I'm going to start from scratch with the processing. Guess I'm pushing my luck finding the dust now this target is heading in the direction of Scotland's capital city then!

Sent from my iPhone using Tapatalk

[Sp@ce_d]

Had another crack at it from scratch now I know how to get round stars. Interestingly, DBE behaved quite different this time round. Found some dust lurking, so I'll try & get more data if these clouds ever part, although the Moons gonna be in the way till Xmas now I guess. Ah well, always next year..

So it's work in progress. Still like better colour saturation in the stars & need to sort out that gradient, separate it from the dust & that black bar down the right. Anyway, eyes are tired so here's V2 for now..

Cheers