Another M31. 8hr autosave added.

[wimvb]

My take on this image. Basically Gnomus' process.

HSV repair script before stretching

A mild dose of HDRMultiscaleTransform and DarkStructureEnhance script at 20% strength.

No MLT structure enhance.

MLT noise reduction on luminance, and ACDNR noise reduction on chrominance, both with a luminance mask

Crop/rotation of the galaxy.

[gorann]

I was curious to see if there were any more details to get out of the core, so I used Gnomus version and darkened the centre using curves in a separate layer in PS. Then I made a new layer and run it through a high-pass filter (set to 50) to increase the contrast (only using it for the centre and a few Ha spots in the arms), and finally I had a run of Local Contrast Enhancement for the core of the galaxy (Astronomy Tools actions in PS - aka Noel's actions). So all done in PS by using layers so I could do it selectively (like Olly recommends). I think I got a little bit more structures out.

Cheers

[StuartJPP]

I think this is the best so far Göran.

I always advocate people spend money on PixInsight or similar before spending a few hundred more on a "better" scope or camera or etc. Until you can process your captured photons properly (we are learning all the time though) I find it doesn't make economic sense.

Once you know you have reached saturation point with the collected photons using the kit you possess the decision can be made to upgrade actual imaging kit.

Well those are my thoughts anyway...

[gorann]

Another thing you can do is to reduce the stars a bit to make the galaxy stand out. Here I used the PS filter from RC-Astro called star shrink. Also in a layer so I could take out the shrink effect on stars that looked odd afterwards. I think there is nothing wrong with reducing stars (everyone does it) since the only reason they look big is that their photons have leaked into adjacent wells (pixels) on your chip.

EDIT: I now also realized (from looking at Wimvb's image) that there is more dust to bring out around the galaxy, so I also did that using curves in PS

Cheers

 

[MARS1960]

12 hours ago, gorann said:

Another thing you can do is to reduce the stars a bit to make the galaxy stand out. Here I used the PS filter from RC-Astro called star shrink. Also in a layer so I could take out the shrink effect on stars that looked odd afterwards. I think there is nothing wrong with reducing stars (everyone does it) since the only reason they look big is that their photons have leaked into adjacent wells (pixels) on your chip.

EDIT: I now also realized (from looking at Wimvb's image) that there is more dust to bring out around the galaxy, so I also did that using curves in PS

Cheers

 

Wow, now this is definitely the best so far, thanks Gorann.

[ollypenrice]

This is going really well! I think that, if you search the net, you'll find that hardly anyone gets significantly deeper into the core than this. It might be possible to increase further the contrast between the features seen near the core and the core itself but actually finding anything new in there is a tall order. Hats off to the OP and all the contributers.

Olly

[StargeezerTim]

I wanted to have a go, especially as moon and clouds have meant its a no go area for imaging here, as has my busy life!  Havn't got the same nice blues as the others but I tried...

 

[ollypenrice]

7 minutes ago, StargeezerTim said:

I wanted to have a go, especially as moon and clouds have meant its a no go area for imaging here, as has my busy life!  Havn't got the same nice blues as the others but I tried...

 

Measure that colour balance in the background sky...

Olly

[StargeezerTim]

21 minutes ago, ollypenrice said:

Measure that colour balance in the background sky...

Olly

Green was a bit high. RGB is now 23,25 and 24. What do you aim for Olly? If I have RGB balanced numerically (e.g. 24,24,24) it has a strong Magenta cast.

[ollypenrice]

11 hours ago, StargeezerTim said:

Green was a bit high. RGB is now 23,25 and 24. What do you aim for Olly? If I have RGB balanced numerically (e.g. 24,24,24) it has a strong Magenta cast.

I thought the green was very high. I usually aim for 23/23/23 but for some reason I find it difficult with my Sony chip data to keep it as high as that. It's easy with the Kodaks.

I'd say that if an equal value in RGB looks magenta on your screen it might just be your screen but I'm not an expert on this.

Olly

[StargeezerTim]

9 minutes ago, ollypenrice said:

 

I'd say that if an equal value in RGB looks magenta on your screen it might just be your screen but I'm not an expert on this.

Olly

It could be my eyes need calibrating...

[gorann]

On 9/4/2017 at 10:50, MARS1960 said:

Wow, now this is definitely the best so far, thanks Gorann.

Thanks!

You have really good data there and Gnomus did a very good job at equalizing the background (probably the hardest part). I merely polished it a bit.

Cheers

[MARS1960]

12 hours ago, StargeezerTim said:

I wanted to have a go, especially as moon and clouds have meant its a no go area for imaging here, as has my busy life!  Havn't got the same nice blues as the others but I tried...

 

Very nice Tim, still looks like i did mine blindfold .

[StargeezerTim]

On 01/09/2017 at 15:59, gnomus said:

For what it's worth, here is my attempt at your data.  It's not as good as Stu's, but it demonstrates that the data is reasonable, and worth persisting with.  There were some awkward gradients which I have probably definitely rushed a little, and I do wonder if this was shot during a time when the moon was up.  

So this was:

• Split RGB channels in PI
• DBE channels separately
• Linear Fit to Red (probably wrong one)
• Recombine using Channel Combination
• Run DBE again!  (Those damn gradients)
• HT Stretching (not too much)
• Bit of saturation using CurvesTransformation
• Some minor saturation tweaks in PS
• Star reduction (not too aggresive) in PI again
• Range Mask minus Star Mask - then used to provide a little MLT sharpening to structure (not stars) in the galaxy

The problem with processing someone else's data is that you don't feel real 'ownership' of it, and tend to rush things.  This took ~30 mins.  Well that's my excuse anyway!

 

Hi Mr Gnomus...

I have tried to process the same file as you with PS and DBE from PI Light and don't seem to be able to get any of the blues you have. I managed to get a fairly flat background using DBE on the separate RGB channels but when I use Curves and Levels and Saturation etc, there doesn't seem to be anything like the colour balance you have obtained. I only use PI for DBE. If you were just using PS, how would you go about trying to obtain the same colours? I attach a copy of the recombined image, post DBE and have tried processing it as a single RGB image and also with a luminence layer like Stuart did... both with no success whatsoever.  Any tips will be most welcome... Tim. 

Here is the recombined image before any stretching or colour work ... any tips of how PS can be used to replicate your colours gratefully received..

 

[gnomus]

Hello Mr Tim

It is difficult to say how you got to where you did with your merge of 'B' and 'C' - and I very much doubt that it can be recovered from where you are.  However, a quick peek in PI shows what is amiss:

[StargeezerTim]

34 minutes ago, gnomus said:

Hello Mr Tim

It is difficult to say how you got to where you did with your merge of 'B' and 'C' - and I very much doubt that it can be recovered from where you are.  However, a quick peek in PI shows what is amiss:

Thanks Steve, it seems PS didnt combine the channels properly. I'll go back and try again.:-)

[gnomus]

2 minutes ago, StargeezerTim said:

Thanks Steve, it seems PS didnt combine the channels properly. I'll go back and try again.:-)

You could do that, but if you were following my workflow, I would suggest that, after you run Linear Fit, you recombine the channels in PI using ChannelCombination.  Then you have your starting RGB (it might need some tweaking like another DBE and some SCNR).

[StargeezerTim]

2 hours ago, gnomus said:

You could do that, but if you were following my workflow, I would suggest that, after you run Linear Fit, you recombine the channels in PI using ChannelCombination.  Then you have your starting RGB (it might need some tweaking like another DBE and some SCNR).

Thanks Steve, I did manage a reasonable image after using DBE to create flat fields for each channel but subtracting them in PS rather than PI. I probably would find it difficult to replicate the process but I remember using the blue channel curve in adobe Raw filter that seemed to make a difference. Pleased with the final version anyway. It lacks your finesse but at least I found some blue! I noticed the red channel of the image had a lot of streaking and noise that has affected the quality of the background. Great thread this, thanks for the help.

 

[gorann]

1 hour ago, StargeezerTim said:

Thanks Steve, I did manage a reasonable image after using DBE to create flat fields for each channel but subtracting them in PS rather than PI. I probably would find it difficult to replicate the process but I remember using the blue channel curve in adobe Raw filter that seemed to make a difference. Pleased with the final version anyway. It lacks your finesse but at least I found some blue! I noticed the red channel of the image had a lot of streaking and noise that has affected the quality of the background. Great thread this, thanks for the help.

 

Starts looking really good Tim! ...although it is the first M31 I have seen with a blue core so yo may have overdone the blue