NGC 253 - the 2014 version

[pvelez (Pete)]

Here is my 2014 version of NGC 253.

There is plenty more data here than last year's effort. That said, I'm struggling still with PI to get a crisp result. ATW seems to make sharp stars with feathery halos and I'm not sure how to wind back the sharpness while keeping a crisper edge.

Comments are most welcome.

A high res version is here:

http://www.pbase.com/equitius/image/157584885/original

Pete

[RickS (Rick)]

Nice detail, Pete. I'd tweak the colours around and up a bit but that's me I often sharpen the stars and galaxy or nebula details separately using masks. They will usually benefit from different settings.

Cheers,
Rick.

[SimmoW (SIMON)]

Lovely,spectacular image. It looks a bit noisy for a dedicated CCD? Maybe due to processing? Still, a great image.

[Shiraz (Ray)]

nice one Pete. FWIW, I prefer the scale of the JPEG - it seems to hold together much better at the smaller scale. Maybe the seeing was not quite good enough to support the larger size? In any event, there is a lot of detail there.

[DJT (David)]

Really nice image Pete. Making the most of that setup I see.

For me the stars are just a bit blown. Maybe a masked stretch with a bit of Local Histogram enhancement?

Can't comment on the detail in terms of processing though. Still got L plates on.

[pvelez (Pete)]

Originally Posted by RickS

Nice detail, Pete. I'd tweak the colours around and up a bit but that's me I often sharpen the stars and galaxy or nebula details separately using masks. They will usually benefit from different settings.

Cheers,
Rick.

Thanks Rick

I usually receive comments suggesting I've been too heavy with the saturation so I toned it down a bit. I'll take your post as an invitation.

I need to go back to the drawing board with sharpening. Each time I sharpen up the stars, the background goes all blotchy and I have grubby looking halos around the stars.

I'd best watch hose videos again.

Originally Posted by SimmoW

Lovely,spectacular image. It looks a bit noisy for a dedicated CCD? Maybe due to processing? Still, a great image.

Yes, its the processing that adds the noise - sometimes I feel like one of those hamfisted restorers of Renaissance paintings - no matter what I do, it seems to degrade the original.

Thanks for the comment. I think it is one of my favourite targets.

Originally Posted by Shiraz

nice one Pete. FWIW, I prefer the scale of the JPEG - it seems to hold together much better at the smaller scale. Maybe the seeing was not quite good enough to support the larger size? In any event, there is a lot of detail there.

Cheers Ray - the JPEG compression hides a multitude of sins. I think this is another waiting for a repro come Full Moon.

The seeing was so so. The data was taken over about a week. I set my rejection FWHM at 3.3 arc seconds. I could drop it a bit more but then I'd lose half my data. Its the usual trade off.

Originally Posted by DJT

Really nice image Pete. Making the most of that setup I see.

For me the stars are just a bit blown. Maybe a masked stretch with a bit of Local Histogram enhancement?

Can't comment on the detail in terms of processing though. Still got L plates on.

Thanks D

The set up is great - reliable and, importantly, available. I also have my PI L plates on. Can you explain the process you suggest with a masked stretch? I'll take any tips I can get

Pete

[RickS (Rick)]

Originally Posted by pvelez

I need to go back to the drawing board with sharpening. Each time I sharpen up the stars, the background goes all blotchy and I have grubby looking halos around the stars.

You should make sure you protect the dark areas and only sharpen the bright parts of the image, Pete. The easiest way is to make a copy of the Luminance or Lightness and black clip it to make a mask.

Originally Posted by pvelez

Can you explain the process you suggest with a masked stretch? I'll take any tips I can get

MaskedStretch is a PI process that stretches an image iteratively but uses a mask to protect the bright areas of the image and prevent them saturating. It can be used as a substitute for HistogramTransformation (or can be used to do a mild stretch followed by HT to complete the stretch.)

Cheers,
Rick.

[DJT (David)]

Hi Pete

Harry's Tutorials. Last video on this page

http://harrysastroshed.com/pixinsigh...homeinter.html

[loc46south (Geoffrey)]

Hi - I hope you don't mind - I downloaded your image and had a bit of a play in Pixinsight. I think it looks a little bit better. The steps I took:
DBE - to clean up the back ground
Masked Morphilociacal transform - to sharpen up and tone down areas
Curves
Masked Saturation Boost
Reset Histogram.
I'm still very much a learner with Pixinsight -so there are probably 100 other ways to achieve the same result. Great image nice telescope the CDK.

Cheers
Geof Wingham

[pvelez (Pete)]

The thread demonstrates the value of forums like this.

I've learnt 2 lessons today from the comments posted.

Firstly, I've started playing with the Masked Stretch tool in PI - what a great feature! I have found in the past that applying a STF stretch and moving it to HT, bright features are blown out. This works well to keep it under control.

I have so much to learn about processing with PI! But its good to ave some tips.

The other comment from Dave (and I think also Ray) was that the stars looked a bit blown. I've tracked it down to my focussing.

I have the Hendrick electronic focussed on the PW which I operate through FocusMax. As I have a permanent set up, I leave everything on - PC, focuser, mount, camera etc. When I run a VCurve set in FocusMax and then focus, peak focus was initially around 4,000 microns out on the focussed. I noticed that the focuser was last night was reaching focus at about 7,000 microns. That looked odd. So I reset the focuser ie ran it all the way in, reset that as 0 microns and focussed again - focus was just under 4,000.

So what is happening here? I suspect that the focuser loses its way over time - so that the place it believes it is located at doesn't match the actual position. So over time, the focus drifts.

After rehoming the focuser, I found the peak FWHM for my images dropped from around 3.5 arc seconds to about 2.9. Not a huge amount but it makes for sharper stars.

I'd be interested in hearing if anyone else has experienced this. Will send an email off to PW too.

Pete

[pvelez (Pete)]

Originally Posted by DJT

Hi Pete

Harry's Tutorials. Last video on this page

http://harrysastroshed.com/pixinsigh...homeinter.html

Thanks Dave - I had a play with this earlier today. I'll give it a go on this image shortly.

Originally Posted by loc46south

Hi - I hope you don't mind - I downloaded your image and had a bit of a play in Pixinsight. I think it looks a little bit better. The steps I took:
DBE - to clean up the back ground
Masked Morphilociacal transform - to sharpen up and tone down areas
Curves
Masked Saturation Boost
Reset Histogram.
I'm still very much a learner with Pixinsight -so there are probably 100 other ways to achieve the same result. Great image nice telescope the CDK.

Cheers
Geof Wingham

No problem at all Geoff - if the data files were't so huge, I'd post my masters to see how others processed my data.

It looks much crisper - thanks

Pete

[Ross G]

Nice looking galaxy photo Pete.

Good luck with the setup. You're at a great location.

Ross.

[pvelez (Pete)]

I spent some time this working learning how to drive TGV Denoise in conjunction with ACDNR in PI.

Here is the repro - still plenty of room for improvement but I'm much happier with the background on this now

Pete

[alpal]

Originally Posted by pvelez

I spent some time this working learning how to drive TGV Denoise in conjunction with ACDNR in PI.

Here is the repro - still plenty of room for improvement but I'm much happier with the background on this now

Pete

It's looking good Pete.
I think I would have masked those stars from the beginning of processing - they look a little hardened -
& maybe I would have adjusted the colour to get more pink areas on the galaxy -
although that might affect the nice blues you have.
It's hard to say without having the original 32 bit stacked FITS.
Still 100 people would give you 100 different results so who can say?
You certainly have great data which is to be expected with such a premium scope.
Maybe something like this (from your data)

cheers
Allan