Problems with Alnitak

[Xiga]

Ok my method above doesn't really work on Huuuuge stars such as this. But here's another way of dealing with them, and it's extremely simple as well.

You can use the Liquify Filter in Photoshop. Once you open the filter, click on the small 'Pucker' icon in the top left. Then adjust the size of the brush to make sure it encapsulates the star and most of the surrounding halo, and very gently click the mouse once at a time. Each click will shrink the highlighted area towards the centre point. It's not perfect, but for a quick fix it's certainly not bad! See a comparison below, which took me all of 1 minute to do in PS.

ps - I should say, the methods outlined above in relation to blending the under and over exposed images together is the best approach. This is just a simple quick fix in case you have difficulty doing it the proper way.

[alan potts]

4 hours ago, Xiga said:

Ok my method above doesn't really work on Huuuuge stars such as this. But here's another way of dealing with them, and it's extremely simple as well.

You can use the Liquify Filter in Photoshop. Once you open the filter, click on the small 'Pucker' icon in the top left. Then adjust the size of the brush to make sure it encapsulates the star and most of the surrounding halo, and very gently click the mouse once at a time. Each click will shrink the highlighted area towards the centre point. It's not perfect, but for a quick fix it's certainly not bad! See a comparison below, which took me all of 1 minute to do in PS.

ps - I should say, the methods outlined above in relation to blending the under and over exposed images together is the best approach. This is just a simple quick fix in case you have difficulty doing it the proper way.

I will give that a try there is a big improvement in the star but I intend to learn the other way as I see other uses for this method. Many thanks, Alan

[michael.h.f.wilkinson]

Interesting discussion. I am thinking about developing a specialized image filter to deal with problems like these. It might be possible to use so-called morphological attribute filters in new ways to stretch the image selectively, by separating out the stars (even big ones) using a variant of the so called clean algorithm combined with a power levelling. In theory, this should deal with the diffraction patterns around stars, and allow you to stretch the extended objects in various ways, and then recombining them with the stars later. Not too different from the approaches proposed here, but it might be more automatic.

[alan potts]

2 hours ago, michael.h.f.wilkinson said:

Interesting discussion. I am thinking about developing a specialized image filter to deal with problems like these. It might be possible to use so-called morphological attribute filters in new ways to stretch the image selectively, by separating out the stars (even big ones) using a variant of the so called clean algorithm combined with a power levelling. In theory, this should deal with the diffraction patterns around stars, and allow you to stretch the extended objects in various ways, and then recombining them with the stars later. Not too different from the approaches proposed here, but it might be more automatic.

I'm impress Michael even though I don't really know what your talking about. Iam trying the ways to do that have been offered but have to say I am struggling, I either miss something or when I do something right i'm then not sure how i did it.

Alan 

[Stub Mandrel]

I thought I would do an example to show the sort of curve I used to adjust the image and keep Alnitak where it belongs with Gaz and Leccy (i.e. under control...)

The images are:

My original output from DSS, probably after a gentle 32-bit pre-stretch in DSS and gradient exterminator.

The curve I applied

The final image after just the curve and then movement of the black point. You can see the appearance of Alnitak has barely changed but the nebulosity has been brought out. I can't help thinking this is on the way to a much better result than my process in November - less is more! (I would stress that an advantage of this curves only approach is retaining the appearance of Alnitak';s binary companion).

[gorann]

On 1/23/2018 at 00:49, Xiga said:

Ok my method above doesn't really work on Huuuuge stars such as this. But here's another way of dealing with them, and it's extremely simple as well.

You can use the Liquify Filter in Photoshop. Once you open the filter, click on the small 'Pucker' icon in the top left. Then adjust the size of the brush to make sure it encapsulates the star and most of the surrounding halo, and very gently click the mouse once at a time. Each click will shrink the highlighted area towards the centre point. It's not perfect, but for a quick fix it's certainly not bad! See a comparison below, which took me all of 1 minute to do in PS.

ps - I should say, the methods outlined above in relation to blending the under and over exposed images together is the best approach. This is just a simple quick fix in case you have difficulty doing it the proper way.

 

That Liquify Filter was a new one to me - seems to be very useful for shrinking specific stars! (although it will of course not bring out the small double star in Alnitak).

Alan, do you see the double star in you unstretched image?

[Xiga]

5 hours ago, gorann said:

That Liquify Filter was a new one to me - seems to be very useful for shrinking specific stars! (although it will of course not bring out the small double star in Alnitak).

Alan, do you see the double star in you unstretched image?

I agree Goran. I quite like it myself, but it's definitely just something you would use sparingly on the odd star here and there, rather than on a whole image.

[alan potts]

21 hours ago, gorann said:

That Liquify Filter was a new one to me - seems to be very useful for shrinking specific stars! (although it will of course not bring out the small double star in Alnitak).

Alan, do you see the double star in you unstretched image?

Yes indeed I do see the double, it is the stretch that is killing it. I feel this data whilst not that bad was taken on an evening when the seeing was not it's best there was also a slight haze , Other nights, I am only learning at the moment so I have had a few goes at it, the star seems better controlled and generally tigher. I am learning more all the time in PS but I can't pick up this method of Robs (U235) that is clearly a good way as it is just too quick and I can't see what is being done.

Alan

[ollypenrice]

Here's an easy way to reduce big stars. It might not be up to getting Alnitak dead right but it should help.

Make a copy layer and, with the top layer active, take an eraser brush, fully feathered (Hardness slider to the left) which comfortably covers all the bloat. Apply it 100% in one click centered over Alnitak. Nothing will change because the top layer is the same as the bottom layer. Choose another cursor before you accidentally erase anything else!

Now make the bottom layer active and open Curves. Alt click with the cursor placed on the brightest real nebulosity just next to Alnitak. This will create a fixed point on the Curve (which is a straight line at this stage..) Put one or two anchor points on the curve below the first one. Be sure that their input and output values are the same so they just hold the line in place. Now get hold of the curve above the highest point and pull it gently downwards. With some manipulation you should be able to get the brightest parts of Alnitak down while creating no artefacts. Because both layers are visible you can see in real time what the result image will look like. When you like it, flatten. Doing this in several gentle iterations with progressively smaller brushes often works well.

Olly

[alan potts]

Thanks Olly, I will write this down and try it out at some point. my worry is that I am trying to many new techniques and I am getting mixed up, you know this is a steep learning curve and little by little is the only way for me but I will try it and thank you for taking the time to tell me.

Having written it into steps it seems fairly easy, famous last words.

Alan.

 

[alan potts]

I rushed this a bit as it's lunch here but wanted to at least try Olly's method of Alnitak taming. I do not believe for one moment I have done as well a Olly would do but it does at least show it is a double. Whilst playing as one tends to do when new to AP, I put a Luminence layer over the top and knocked it back to 66%. I am sure I could do better but thought I would let you all see. Imagine trying this in a darkroom years ago.

[Stub Mandrel]

Gone too far, I think, but if you can achieve that you can get anywhere in between.

[alan potts]

8 minutes ago, Stub Mandrel said:

Gone too far, I think, but if you can achieve that you can get anywhere in between.

On reflection i don't disagree with you. I rushed it really but with still having the data as you well know can revisit. I feel the original data could have been better as I have a small amount taken at an earlier date where Alnitak is nowhere near as bad on the same length subs.