Username:

Password:

Altair AstroDIO DehumidifiersAtik CamerasModern AstronomyDavid HindsNe3 Filters
Solar Gravity 2018

Author Topic: Image Processing Challenge #1 part 3 - post processed images  (Read 13640 times)

0 Members and 1 Guest are viewing this topic.

Offline dciobota

  • Daniel
  • Moderator
  • Galactic Poster
  • *****
  • Posts: 9722
 Ok folks, Rob was kind enough to upload Ben's data so it can be downloaded.  Here are the links:

http://ukastroimaging.co.uk/fits/horsehead_LRGB.zip

http://ukastroimaging.co.uk/fits/horsehead_OSC.zip

The second one is the one combined fits file btw.  I'll post a zipped tiff of the one fit when Iget the chance.  In the meantime, those of you who have photoshop can use Fits Liberator to load them into photoshop.

Ok, I created a color 32bit tiff from the one fits file for those who can't use fits liberator. The zip is here:

http://ukastroimaging.co.uk/index.php?ind=downloads&op=entry_view&iden=3

 
 Have fun, and post your submissions here.  :)

Daniel
PLEASE REMOVE ME

Offline roundycat

  • Poster God
  • *****
  • Posts: 1789
    • Dennis Isaacs Astrophotography
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #1 on: 20:33:23, 28 July, 2007 »
Hello all

I can't quite express my amazement that Ben managed to get what he did from that image. My opening move was to measure the colour of Alnitak. 47,254,255 RGB!!!!! Where did all the red go??? First off I righted the image as it is a personal thing, looking at something upside down and inside out does nothing for me! I ended up using levels twice and curves no fewer than five times to get the colour somewhere near what I thought was right. Also included was a selection around Alnitak to dim it and then inverse the selection to brighten the rest. The curves adjustments I used in the red channel were really scary and it did bring up the noise. Neat Image contained that with a little bit more brightening of the main image area after Neat. Then USM at my normal rate of about 200/0.8/12.

My initial reaction was to go and read a book but I am glad I persevered. The finished effort is below.

Dennis

Edit: I have been a bit cheeky. I have changed the picture to v2 as it was done in 'proper' lighting and is an improvement.

[attachment deleted by admin]

Offline jaycee

  • Full Member
  • ***
  • Posts: 331
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #2 on: 22:53:26, 28 July, 2007 »
FWIW... I was not sure what I was doing but this is it, anyway. :D
Thanks for the opportunity to make a fool of myself again. :)

Cheers! wesdasd
Jeff

[attachment deleted by admin]
Jeff wesdasd ~ Celestron Nexstar 102 SLT

Offline dciobota

  • Daniel
  • Moderator
  • Galactic Poster
  • *****
  • Posts: 9722
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #3 on: 07:47:17, 29 July, 2007 »
Great job guys! :cool:  I haven't tried to process yet, but I can tell it's a toughie.  Btw I updated my post with a link to a tiff file made from the second set of fits.  No processing though, I used the original scaling.
PLEASE REMOVE ME

Offline dciobota

  • Daniel
  • Moderator
  • Galactic Poster
  • *****
  • Posts: 9722
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #4 on: 08:38:49, 29 July, 2007 »
 Ok, did a relative quickie on it, but tbh, I don't think I can do any better.  Dennis, awesome work on reducing the star size, all my efforts didn't get me near that.  Jeff, I like the colors in yours, imo all the reflection nebulas in there, including the reflection from Alnitak, get lost sometimes if the Ha overwhelms... but that is just a personal preference.

 Here's my version, my colors aren't nearly as nice and still too washed out I think.  I did a similar procedure to Dennis', I selected a failry large area around Alnitak and did a series of gaussian blurs with fades to darken, to reduce and smooth the blown areas, then a couple of curves to dim it a bit.   The rest was just a couple of deep space noise reduction actions, a bit of smart sharpen and a couple of star reductions. Then I increase star colors a bit and then an overall saturation increase.  That's it really.

 

[attachment deleted by admin]
PLEASE REMOVE ME

Offline roundycat

  • Poster God
  • *****
  • Posts: 1789
    • Dennis Isaacs Astrophotography
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #5 on: 09:25:05, 29 July, 2007 »
Daniel,

this sort of thing is going to be a winner. I thought I might explain a bit more about my processing as it is clear that too short an explanation can lead to all sorts of misunderstandings.

I don't think I mentioned that of the two versions posted I worked on the RGB one.

Most of my curves adjustments were kept small and targeted onto a specific brightness range in the picture. Mouse over the picture with the left button held down and then contol click at the brightest and darkest points in the area of interest. That places adjustment points on the curve exactly where you need them. Then tweak the contrast whilst being careful not to let the top of the curve go anywhere near saturation.

My selection of Alnitak was not an ordinary one. I used quick mask and applied a radial gradient centred on the star. This was then swapped for a selection outline and the selection reduced in size until it only just covered the centre of the star. I then used levels to gently lift the black and grey points to darken the star without it being 'obvious'. I then inverted that selection and brightened the rest of the picture using curves.

You can also 'save as' a partially processed version and subsequently use the whole thing or part of it as a layer mask to darken bloated stars. Layer mask has the advantage of course that it works on brightness rather than area.

The Ha emission below Alnitak is often shown on star maps as a triangular area and I applied quick mask to that area in the shape of a narrow triangle. This was turned into a selection outline and then feathered until the outline disappeared. A warning comes up saying that no pixels were 50% selected and the outline will not be visible. Just what you want! I then used curves again (6th time I think!) to increase the contrast, mainly in the area of the head but also to brighten it all a small amount.

A few final points: I never use Fade as I don't know what it does. I prefer to work on a duplicate layer and then reduce opacity/change blend mode/partially erase/layer mask or selectively enhance, at least it gives you a feeling of being in control.

I never use smart sharpen. My experiments with this ages ago never proved to me that it was any better than USM, if necessary preceded by a small, possibly selected, blur. 

As they say, your mileage may vary :D

Dennis

Offline 661-Pete

  • Poster God
  • *****
  • Posts: 750
  • W.Sussex & SW. France
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #6 on: 09:51:07, 29 July, 2007 »
Yup, apart from anything else, this thread, and the detailed write-ups that have gone into it, is adding an amazing amount of new material to the knowledge base on image processing.  Just to realise how many ways there are to 'skin a cat' is daunting - all I know of the subject is the odd tweak just to make a capture 'postable on UKAI', that's all.   Of the above I already have a 'favourite' but shall keep  :-X till all the entries are in and the voting starts.
Pete

(Orion Europa 250, Vixen GP, Canon EOS350D) Telescope and mount in France, but not doing much at present..

Offline egdavies

  • Galactic Poster
  • ******
  • Posts: 7399
  • Necessity is the mother of all inventions
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #7 on: 11:33:20, 29 July, 2007 »

A few final points: I never use Fade as I don't know what it does. I prefer to work on a duplicate layer and then reduce opacity/change blend mode/partially erase/layer mask or selectively enhance, at least it gives you a feeling of being in control.

I never use smart sharpen. My experiments with this ages ago never proved to me that it was any better than USM, if necessary preceded by a small, possibly selected, blur. 


Hi Dennis,

Smart sharpen is a  type of Deconvolution. With this type of sharpening you are not so prone to getting black/white lines around brighter objects due to the way USM works.

Fade. This works by ignoring any layer effects that you have opn that layer. That is, it does not fade out the layer effects that you have, unilke using opacity which fades out the layer effects also.

Hope this helps.
Eddie
MK1 Eyeball

Offline jaycee

  • Full Member
  • ***
  • Posts: 331
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #8 on: 12:50:34, 29 July, 2007 »
Thanks for the useful comments, Daniel. (/goes away to look up the meanings of Alnitak, Ha and reflection nebula. ;) )

I had tried to use Nebulosity but I am really not familiar with its settings just yet nor am I completely comfortable with what they all do. I used PSCS and the FITS plugin to process the RGB images. Noel's tools to initially bring the images directly into the FITS PSCS plugin, where each image was subjected to stretching with the arcsinh(x) function.

Noel's tools to reduce deep space noise and star size reduction + fade sharpen to mostly lighten and also enhance DSO and reduce stars. Histogram  tweak of the final image... reduce to jpeg image. ta dah! (How not to do it) :D

I feel that a little more contrast around the horse's head would have been helpful. Reducing the size of Alnitak was not something I had considered... something about messing with reality that may be inbuilt into all photographers... although i am inclined to think that adjusting something that we are seeing as it was 817 years ago may not count as altering reality. :D ;)

Cheers!  wesdasd
Jeff
Jeff wesdasd ~ Celestron Nexstar 102 SLT

Offline dciobota

  • Daniel
  • Moderator
  • Galactic Poster
  • *****
  • Posts: 9722
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #9 on: 15:31:26, 29 July, 2007 »
... although i am inclined to think that adjusting something that we are seeing as it was 817 years ago may not count as altering reality. :D ;)

Lol Jeff, I like the way you think! :D  Looks like you're getting the art of astro processing down pretty nicely too.

 I agree guys, like the work done on Mike's Eagle pic, this not only shows the different ways to process but also how subjective the "right" look can be.  This is why I would never make this a competition.

 Dennis, I like the way you used masking on Alnitak and the Ha areas, I may try that myself, looks like you get better control over contrast at least.

 Btw I use fades cause I'm lazy, it shortcuts creating a duplicate layer, changing blend mode and then flattening again.  Smart sharpen is also a shortcut because it can mask out the levels you don't want sharpened.  I think it also uses a different deconvolution algorithm than usm.  Btw Eddie, you can still get halos if you overdo it, that's why I still fade it to lighten.


PLEASE REMOVE ME

Offline egdavies

  • Galactic Poster
  • ******
  • Posts: 7399
  • Necessity is the mother of all inventions
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #10 on: 15:35:54, 29 July, 2007 »
Btw Eddie, you can still get halos if you overdo it, that's why I still fade it to lighten.

I know Dan, its just you can push it more with smart sharpen before you do get them :D

Eddie
MK1 Eyeball

Offline roundycat

  • Poster God
  • *****
  • Posts: 1789
    • Dennis Isaacs Astrophotography
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #11 on: 17:29:54, 29 July, 2007 »
As Pete says above, it is certain we will all learn more about processing, even if it is only what not to do. I have about five books on PS and none of them mention Smart Sharpen although I seem to remember having used it years ago, just to see what it did.

Somebody please tell me where I can find it. I shall try it on an astro picture (never have) but I have to make a point in defence of USM. Ugly halos are only an issue if you oversharpen, ie, too much Amount or too much Radius. The received wisdom in the world of publishing is to set the radius to your final PPI resolution divided by 200. This is based on half tone principles which usually screen at 300ppi for high quality work.

Seeing halos around your stars is a sure sign you have overdone it. If you sharpen for inkjet or other continuos tone device sharpening that is visually pleasing at 100% screen magnification usually works just fine. If you sharpen a full size image as a TIFF and overdo it a bit it will look worse when saved as a jpeg for the web. I am convinced the thing to watch for is to avoid halos by keeping the Radius and particularly the Amount down to sensible values. I read some complete tosh in Astronomy Now this month about starting with 500% Amount and working down from there. 150-200% usually does it for most images. The Radius can be safely left in the 0.7-1.5 pixel range. It seems important to me at least to keep the Threshold up to around 10-15. This avoids most of the tendency to sharpen smoother areas as this turns into posterisation anyway.

The last thing to remember about sharpening is that it is a contrast increasing device. Use the info palette and carefully measure the RGB levels of several stars. You will need to magnify quite a lot to do this. Then sharpen to your normal taste. Then measure those levels again. They will all be higher, so high in fact, that they may all saturate so your bright, coloured stars become like white dinner plates.

My conclusion is that USM is a good tool to whiten all your stars and produce ugly halos if used improperly. There is an escape clause though, as with so much in PS. Duplicate the layer and sharpen it. Then set the blend mode to lighten and if you have made lots of halos they will all be absorbed by the lighter pixels underneath and you will be left with your oversharpened, white stars with no halos at all.

Eddie, one last point. Your mention of the Fade option left me wondering so I looked it up. Quote from PS CS user guide: 'The Fade command changes the opacity and blending mode of any filter................  Applying the Fade command is similar to applying the filter effect on a separate layer and then using the layer opacity and blending mode controls'.

Once again congratulations to Daniel for a good idea. If we can get the submissions this one will run and run,

Dennis

Offline egdavies

  • Galactic Poster
  • ******
  • Posts: 7399
  • Necessity is the mother of all inventions
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #12 on: 17:46:14, 29 July, 2007 »
Doh! Ignore me Dennis, I was thinking of Fill rather than Fade  :ninja:

Smart sharpen is just above USM sharpen! What version of photoshop are you using? It was introduced in CS2.

Wisdom in the world of publishing? Have you worked in publishing? :lol:

Eddie :D
MK1 Eyeball

Offline roundycat

  • Poster God
  • *****
  • Posts: 1789
    • Dennis Isaacs Astrophotography
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #13 on: 17:53:47, 29 July, 2007 »
Eddie

I'm very pleased to be able to report that all my qualifications are in Electrical Engineering and Electronics although I had a 30 year career in BT Procurement specialising n QA and Reliability. I no doubt have the same opinion of the world of publishing as you do. Real World Photoshop however is a real bible for the programme. Written by two guys that have used if since its inception. If you look up Bruce RGB, a strangely named colour space, it is named after one of the authors.

I am still on CS, I am also beginning to think I was using Smart Blur in the past as it could not have been Smart Sharpen. That's why I couldn't find it. (repeat after me, Doh!!).

Dennis

Offline egdavies

  • Galactic Poster
  • ******
  • Posts: 7399
  • Necessity is the mother of all inventions
Re: Image Processing Challenge #1 part 3 - post processed images
« Reply #14 on: 18:06:56, 29 July, 2007 »
DOH!  :big_clap: :lol:

If case you did not know Dennis I work in publishing! In fact I get to play with PS all day! :D

Eddie
MK1 Eyeball

 

ukbuysellRemote Imaging from AustraliaSharpSkyblank APTUKAI on Facebook
Powered by SMF 2.0.15 | SMF © 2006, Simple Machines LLC
DarkBreak by DzinerStudio. Theme modified by The UKAI Team

Page created in 0.706 seconds with 38 queries.
TinyPortal © 2005-2012