Painting a mask with the cursor

[LX200]

Has this ever been considered... an ability to "paint" over parts to the shot to create a mask?

[msmythers]

You might want to read this posting.

http://pixinsight.com/forum/index.php?topic=5624.msg38543#msg38543


Mike

[LX200]

Bah... it's ridiculous to suggest that if you create a complex mask through a series of PixelMath gymnastics, that this is science, but if you paint over a nebula to isolate it for processing, this is a corruption of the highest order.

[Herbert_W]

+1

[troypiggo]

You can use the CloneStamp process.  The source image does not have to be the image you're cloning onto, so use an all-white or all-black image of same size as the one you want to paint onto.

[Juan Conejero]

it's ridiculous to suggest that if you create a complex mask through a series of PixelMath gymnastics, that this is science

Yes. The problem is not in how you paint the image, but in the fact that you decide to paint the image. There can be special cases where removing certain features from a mask, based on objective criteria, can be acceptable (see an example here). With these exceptions, which have to be analyzed case by case, and have to be based on a previous analysis to identify and solve specific problems, there is no reason to paint an image or a mask in astrophotography. It is not about science versus art; it is all about the documentary nature of astrophotography.

You can use the CloneStamp process

CloneStamp has been designed to fix small defects that can be identified without ambiguity, as far as there is no risk to remove significant features. It is not intended to replace brushes, pencils, lassos and similar hand painting tools. If you want to use these, there are many painting and retouching applications that are much more suitable than PixInsight to perform arbitrary manipulations.

[LX200]

Well, here is where my frustration stems from.... I have a shot of M27, and it of course shows thousands of stars through the entire field.  I wanted to tweek the color a bit on the nebula part, and add saturation, without effecting the stars....

So... I tried Luminance mask minus star mask to get a mask for just the nebula.  But I cannot, for the life of me, get a mask of the stars that does not create halos, misses stars, etc... because there are just so damn many stars of different sizes, etc.

I just find myself sitting there going "Oy, with a feathered paint brush I could do this in 7 seconds"

[RickS]

Well, here is where my frustration stems from.... I have a shot of M27, and it of course shows thousands of stars through the entire field.  I wanted to tweek the color a bit on the nebula part, and add saturation, without effecting the stars....

So... I tried Luminance mask minus star mask to get a mask for just the nebula.  But I cannot, for the life of me, get a mask of the stars that does not create halos, misses stars, etc... because there are just so damn many stars of different sizes, etc.

I just find myself sitting there going "Oy, with a feathered paint brush I could do this in 7 seconds"

Create a clipped luminance mask, remove the stars (I typically use PixelMath "iif(star_mask>n,0,$T)" with "n" tweaked for best result) then give it a Convolution blur to remove the hard edges.  Voila!

[LX200]

1. I know how to make a luminance mask... How do I make it clipped?
2. What is star_mask in equation... Just a mask made with star mask tool from luminance?
3. What is $T... The image I drag pixel math operator onto...?. The luminance mask?

[RickS]

1. I know how to make a luminance mask... How do I make it clipped?
2. What is star_mask in equation... Just a mask made with star mask tool from luminance?
3. What is $T... The image I drag pixel math operator onto...?. The luminance mask?

1. Use HistogramTransformation and move the shadows clipping point to the right
2. Just an ordinary star mask created with the StarMask process.  It doesn't need to be incredibly precise if you use the PixelMath expression I suggested
3. $T means the target image of the PixelMath process.  You should apply PixelMath to the clipped luminance mask.

Cheers,
Rick.

[llpastro]

Rick, what value of n is a good place to start?

Larry

[ChoJin]

I'd say use the readout mode with your cursor around the star halos to decide

[RickS]

Rick, what value of n is a good place to start?

Hi Larry, the suggestion to look at readout values is a good one.  My lazy approach is just to start at 0.1 and adjust up if the holes around the stars are too big and down if they are too small.

Cheers,
Rick.

[llpastro]

Thanks to both of you for your suggestions.

Larry

[jerryyyyy]

Doing some leisure reading here and never get this to work since I can never make a good star mask with 8000 or so stars in a nebula shot of the Milky Way.  Anyone have some good suggestions for basic star mask settings?