M31 LRGB 19hrs - PCC and ArcsinhStretch

[drmikevt]

Continuing the processing of my big M31 project, which  was very unfortunately taken while I was still having some flatness issues.  It affects some of the outside stars if you look (don't look). 
L : 72x300 + 120x30
RGB : 24x600 each
MyT, SV80ST, QSI 683wsg-8
 
Luckily, I was just about to start this when ArcsinhStretch was released.  It is very easy to use - I like it already.  I compared it to Masked Stretch and Hist.Trans and it clearly provides much better color information.  Color in this image was not touched except for a very, very light curves saturation to the whole image; otherwise, it is exactly what the (also new) PCC tool and ArcsinhStretch produced.  It begs the question -  is is just bringing out color to a more natural extent than the others are able to or is it saturating things beyond what is 'really' there?  I'm not sure, but the results are nice either way.  Please enjoy and I'll take any and all suggestions for improvement or rework.  Of course, the full version, linked below, is much nicer (as always)
 
http://www.astrobin.com/full/316529/0/

Thanks for looking
Mike

[lizarranet]

Hi,

I'm processing also a M31. Roughly 12h with a FSQ106 f3.7 and a Moravian G3-16200. It's my second try after a first attempt some years ago.
The colour balance in my images of M31 is like yours. Im not able to obtain that saturated blue color on the periphery of the galaxy that you can see in many images out there.
I wonder what I do wrong and I do not find the reason....

Regards,

Mikel

[RickS]

Mike,

That looks quite nice.

One small point: I find arcsinh stretch a bit too heavy and have been blending it with a conventional HT stretch and find that looks more natural.  Some of your big orange stars have the look that I'm trying to avoid with a saturated blobby star with a hard core.  They look a bit like a pimple    It's not a big deal  but I think there is a small improvement you could make.

Cheers,
Rick.

[sharkmelley]

Great image!  It's come out very well.

Luckily, I was just about to start this when ArcsinhStretch was released.  It is very easy to use - I like it already.  I compared it to Masked Stretch and Hist.Trans and it clearly provides much better color information.  Color in this image was not touched except for a very, very light curves saturation to the whole image; otherwise, it is exactly what the (also new) PCC tool and ArcsinhStretch produced.  It begs the question -  is is just bringing out color to a more natural extent than the others are able to or is it saturating things beyond what is 'really' there? 

That's a good question about ArcsinhStretch.  The answer is that it is simply showing the colour that already exists in your image, without manipulating it in any way.

The philosophy behind it is that within each pixel the original proportions of R, G & B are preserved during the stretch.  It is only the "brightness" of the pixel is altered.  Stretching the data using alternative methods usually results in the pixel's colour being bleached, especially the brighter pixels.  Such methods are usually followed by some kind of colour saturation operation to restore colour.

I agree with Rick's comment that brighter stars do tend to have a certain look about them following a powerful ArcsinhStretch operation and so it should be ideally combined with another process to address this.

Mark 

[drmikevt]

Hi,

I'm processing also a M31. Roughly 12h with a FSQ106 f3.7 and a Moravian G3-16200. It's my second try after a first attempt some years ago.
The colour balance in my images of M31 is like yours. Im not able to obtain that saturated blue color on the periphery of the galaxy that you can see in many images out there.
I wonder what I do wrong and I do not find the reason....

Regards,

Mikel

Mikel - I think that the saturated blue that you see in many images is a fabricated color effect to make a pretty image.  The goal of this image was to try to present a 'natural' color profile, as determined by PCC, without altering the colors too much.  Its not that those other images are 'wrong', its just that that heavy blue does not really reflect reality (or, what we would see if we were floating in space).  As it is, this image is probably a bit saturated compared to what we would see in space, but I'm good with that.  In the end, make the image that you like.  If you don't already know this trick, to boost color, you can try making the areas you are interested in and then use MLT, slightly increasing bias (.05-1.5) on layers 5-8 with it set to target Chrominance. Then, saturate slightly - that should help bring out your blues. 

Mike

[drmikevt]

Mike,

That looks quite nice.

One small point: I find arcsinh stretch a bit too heavy and have been blending it with a conventional HT stretch and find that looks more natural.  Some of your big orange stars have the look that I'm trying to avoid with a saturated blobby star with a hard core.  They look a bit like a pimple    It's not a big deal  but I think there is a small improvement you could make.

Cheers,
Rick.

Thank you, and thank you for the tip!  I do think that will improve things.  I also used the HSV Repair script before stretching and then did a slightly Convolution to the big star cores to deal with some clipped pixels after AS - I think things would have been even worse without those (but maybe the convolution is contributing to what you are seeing as a problem).  Plus, my stars are not yet as crisp as I'd like (I think my reducer is at play, but not sure yet) so that doesn't help either. 

Mike