Narrowband color composition using J-P Metsavainio Tone Map technique in PI

[astroedo]

Hi all,
I've just published on my website a new tutorial on narrowband composition.

It is an adaptation of original J-P Metsavainio technique.

I hope you enjoy it

http://www.radice.biz/articoli/translations/75-narrowband-color-composition-eng

Bye

[oldwexi]

Hi Astroedo!

Excellent work!!!!!
You got me also a new idea  about how to create good starmasks for star reduction!
Thanks a lot for making this available.

Gerald

[astroedo]

Great! I can't wait for it

I also have an Idea for unconventional starmask building:

I'm experimenting with PixelMath to create a "clipped gradient mask" to use as a Deconvolution support.

I will post as soon as possible

[kolec]

Very interesting . Thank for inspiration.

kolec

[RickS]

Thanks from me too, Edoardo.  That was an interesting read.  The last couple of NB images I processed I did something quite similar except that I didn't remove the stars completely - I just reduced the size of the OIII and SII stars significantly before I did the combine to generate the colour component of the image.

Cheers,
Rick.

[jkmorse]

Edoardo,

Thanks for posting!  Fascinating read.  Been struggling with some really good NB data of NGC281 and looking forward to trying your technique.

Best,

Jim 

[Geoff]

Very nice Edoardo.  The main disadvantage is that you lose all colour in the stars.  Do you see any way around this?
Geoff

[Geoff]

I tried this process to get rid of the stars on an image, just to see how it went.  It seems that the star mask is critical.  When I applied MMT after applying the star mask to the image, it got rid of most of the stars, but not the very little ones nor the very bright ones. Some more tweaking got rid of a few more of the bright stars but still not the little ones--I suppose I have to experiment a bit more (a la Harry).
Geoff

[astroedo]

Thank you Geoff!

Yes the star mask is crucial, sometime I use two different settings to get all the stars, and I use a range mask enlarged with morphological transformations for the big saturated stars.

Then I merge the different masks using the max() operator in PixelMath.

I've Just downloaded Tone Mapping V2 by Metsavainio where it explains how to recover stars colors and I'm working to integrate it in the normal workflow in PI.

Unfortunately I don't have so much time as I would like to make tests and write tutorials... 

[jkmorse]

Edoardo,

A quick question about the boosting factor formula you present for Hubble Palette color balance.  What is the point of the second half of the red and blue formulae ("-(OFFSET*(SII[OIII]Boost-1))").  Unless I am missing something, the deduct, in the Sii case, takes the boost of 7 down to 6.997 and in the case of Oiii, reduces the boost from 4 to 3.9985.   

Do those miniscule reductions really add anything to the result, can they just be ignored, or am I missing something?

Thanks,

Jim

[astroedo]

Its goal is to keep the background with a neutral color.
without this correction you would have a strongly colored background because its value would be boosted.

[jkmorse]

Thanks for the clarification.

Best,

Jim

[Buzz]

The tutorial is great, many thanks - I'm trying to achieve the blue and gold motif (traditionally by using selective color adjustments in Photoshop- removing cyan from green for instance). In your tutorial, even with scaling, I'm getting a very green result from the dominant Ha. Are you using some wonderful curves transformation to achieve the color balance in the last two steps?

[astroedo]

Hi, I sometimes use InterChannelCurves Module using Green (or another color channel) as Both reference and target channel.

InterChannelCurves is an optional module that you can downolad for free from Carlos Milovic website http://www.astrophoto.cl/Research.html

you can also find information here.

http://pixinsight.com/forum/index.php?topic=2275.0

I hope this helps 

[troypiggo]

Have you tried Rick's new script?
http://pixinsight.com/forum/index.php?topic=7751.0