Double Cluster for comment please

[dlp]

With the cloud settled in forever, I've been spending my evenings trying to learn PixInsight. Harry's videos have been an inspiration!

I decided to have a go at some images I took of the DC with my Nikon in November, but never got round to properly processing the raw images rather than some resized tifs. So here we are, 15 subs, 10s - 20s ISO 1600, darks & bias (flats were wrong size d'oh!). My aim was to properly get to grips with DBE, star masks, ACDNR using the inverted star mask, & selective colour enhancement.

I've included the whole frame (although I can see some slight elongation of stars in the corners despite using a coma corrector) as it looks nicer than a crop. It's obviously been reduced from the original 4928x3264px.

Using PI has prompted me to order a doubling in RAM to 8Gb as dealing with this size image is a bit boring on the processing wait time.

So, please tell me what you think, not enough colour? too blue? too black? I'll use any feedback to lock myself away for another week of evenings...

David

Edit: I can see I could have probably sharpened a little more for web use.

[Nexus 6]

Am no imager David, looks a nice shot of one of my favourite sights! If you pushed me I would say you had a tad too much blue, I am sure you will have some more concrete opinions from the imagers on the forum.

Alan

[Pete Presland]

so would i,but really nice image anyway! well done

[oldfruit]

Nice image and some good colours too.

[ollypenrice]

Too blue, yes, though the red stars are showing nicely. Doesn't that fine Mr Page have a tutorial on the 'all stars' method of colour balance? Also have you tried background neutralization for getting a black sky?

I can give you a nifty star rounding technique for Ps if you need to doctor your corner stars regularly. I do this to avoid spending 600 dollars on a flattener for our TEC! The truth is coming out... Oh dear!

Olly

[R Fisher]

Nice image, Star colour's are definatly there.

On my monitor the background of the attachment image looks quite dark, may be worth trying to lift it a little.

If there is a Blue cast in the background (not sure) I would try 'Background Neutralization' in PI.

[dlp]

Thanks for the useful comments Guys. Just been having another play (8Gb RAM arrived, woop!). I've reduced the blue on the highlights and in the process made a better star mask. The b/g was fairly neutral but I applied 'Background neutralization' again just to be sure. It's not 100% black, but I notice the process of jpeging does darken it slightly.

I'm getting to the stage now where there's very little to do (in fact can do) in PS once an image comes out of PI, though I do like PS's 'smart sharpen' feature, and selective adjustments are easier too.

What I don't like about my image is its flatness (ie doesn't look so wonderfully 3D as it does when you look through the scope). I'll maybe try some of Ken Crawford's techniques to improve it although I guess a lot more subs would be the biggest gain.

And yes please Olly, would love to see your tut on rounding stars selectively!

David

[jgs001]

Looks very nice David, the second has better colour to my eyes... I'd love to hear Olly's trick too...

[harry page]

Hi

Good star me chum

The background is neutral as well

A couple of things to help

1) I think it could stand more of a stretch to make the stars pop

2) you could use the all stars c calibration ( use the whole image )

3) There is also the assisted colour calibration you could try ( like a channel mixer

4) use background neut after 3

Also I have just updated the DBE video to show a little more info

Regards Harry

[Tim]

Hi David.

Its not too shabby at all

To my eyes, (albeit slightly inebriated and without my glasses on) the stars look a tad soft, was the focus perfect?

The other thing, easier to see with the whole pic and not a crop, is whether collimation is spot on? Are the distorted stars evenly spaced and showing the same distortion all around the image, allowing for it being rectangular of course? I only ask because the stars seem stretched in an odd direction for coma alone, there might be something else at play.

Cheers

Tim

[harry page]

Hi

Oh I forgot to say " the sky is never black "

Harry

[dlp]

TJ - may well have been a collimation issue as I had messed about with the secondary having fitted a new focuser, hopefully now rectified. Well spotted. Focus may well have not been perfect as I was using trial and error with exposures until I learned to use the live display on a bright star.

Harry, thanks for your useful comments too, and I was pleased that I had played with the techniques you mentioned. Had a look at your new DBE vid too - very good indeed particularly on fixing the dreaded red samples! I was a trifle disappointed that you didn't remove the remaining gradient in the first image though

And thanks everyone else for pitching in, it's been really helpful.

I've attached another stab at the image, but I've probably over-egged the highlights this time, hey ho another couple of hours gone...

Blacks are not clipped but I took them down to remove some unpleasantness going on in the b/g.

David

[harry page]

Hi

Colour is more restrained this time perhapes a little sat boost might help

Sorry about the dodgey video , will go and beat myself with a stick

Harry

[Tim]

Do you have a Bahtinov mask for the telescope? They take all the guess work out of focussing, and enable perfect focus every time.

Many (too many!) times I have tried to work through data that wasn't quite up to the job, better sometimes to start again, but you can still learn the procedures with them. Pictures that features stars as the target, like the double cluster, absolutely must be focussed and nicely collimated to get the best from them. It's not quite as bad as the 'polishing a [removed word]' cliché, but you get the idea

[dlp]

TJ - not made/got one yet; on the list, although I've now found the 'live view' method great, once you have a bright enough star. Will try harder!

And Harry, no offence intended re my comment on the video - I (I'm sure WE all) appreciate the time it takes you to make these. Hope Pleiades are sharing the profits with you!!

Regards your 'increase sats' comment - is that the equivalent of the electrician's 'go get me a bag of sparks from the stores' ?

David

[martin_h]

You could try this plugin for PS star rounder Photos 2

Near thr bottom of the page

[ollypenrice]

My own star rounder action (picked up fropm Peter Vasey) is

- magic wand to select star, then record the following action;

-expand about three pixels

-feather about two pixels

-Filter-blur-radial blur (spin and best quality)

-deselect.

If you have this with a keyboard shortcut (f2 etc) and use the magic wand on the errant star, then hit the shortcut, it is very quick assuming you are not in the middle of the Milky Way!!

You can make the selection bigger and the feather likewise, to suit the kind of spot size you get from your setup.

Olly

[dlp]

Thanks Martin/Olly. Although this image clearly isn't worth it, I've made notes for future reference. I reckon you could probably also make a star mask in PI, export to PS and fill the non-errant area with black, re-import to PI and then use Morphological Transformation using a non-round 'way' (assuming the distortion was linear anyway).

David

[harry page]

Hi

I never take offense mate , Lifes too short

Harry

TJ - not made/got one yet; on the list, although I've now found the 'live view' method great, once you have a bright enough star. Will try harder!

And Harry, no offence intended re my comment on the video - I (I'm sure WE all) appreciate the time it takes you to make these. Hope Pleiades are sharing the profits with you!!

Regards your 'increase sats' comment - is that the equivalent of the electrician's 'go get me a bag of sparks from the stores' ?

David

[dlp]

I had a play with my idea about morphological transformation and distorted stars. It does work. The most difficult bit is finding the right structuring element shape! See below. I'm sure there's some widget hiding away in Pi that already does this, or someone's done it a better way but hey 'experiment' says Harry!

PS Exporting the mask and filling the non-errant areas in PS then re-importing works fine too.

Edit: Of course it works - that's what its intended for. The only slightly novel bit is taking into PS for editing

[harry page]

Hi

There is no star PSF tool in Pi at the moment but it is comming soon

Have you tried decon to sort out the stars , just select a star shape that matches yours and select derringing and go for 4 or 5 iterations

Harry

[dlp]

removed