Problem with extracting natural color from Image

[Andre27]

Good Day Ivo

I am having difficulty in extracting or representing natural color in a particular image , what is given is that I image from a very light polluted site and a a flood light about 25 m from my pier. I use a UHC filter that has proven to be very effective in my other images , I have however had a similar issue with the Rosette Nebula which lays in the same direction as the Wide Field Horsehead/Flame nebula which is the image in discussion.
At first I thought is is just bad data but I do not think that is entirely the case.

The image is 13 hours of integration time / 300 & 480 Sec exposures / darks and flats applied.Not white balanced.
Camera Modded Canon 1100d /ccd temp 29 Degree Celsius ave , Stacked in DSS.

Richard has had a go at it for which I am grateful , however the the result are similar to mine , the links below are for two different stacks of the data , I would appreciate if someone could have stab at it.

Restacked
https://drive.google.com/file/d/1l5i-09 ... sp=sharing
Original Stack
https://drive.google.com/file/d/16_h_Ec ... sp=sharing

Thank you for your time ,appreciated.

Andre

[almcl]

Hi Andre27

I have had a go at your data, although I doubt I could do any better than Richard, result below.

What is it in his (and your) image that you find unsatisfactory? Imaging from a light pollution area is difficult (I have similar situation to the one you describe) and any sort of filter is going to upset the colour balance to some extent, but your data looks pretty reasonable to me and very smooth.

before applying the color module

Combined Flame pre colour.jpg (65 KiB) Viewed 8349 times

after colour and noise reduction

Combined Flame post noise.jpg (82.61 KiB) Viewed 8349 times

[Rkonrad]

Nice job with the colour! But I dont get it. This is my result after obtaining a colour balance by sampling the stars. https://www.flickr.com/gp/rkonrad/3MVS38 . Notice how high saturation setting is. I gave up with the colour and just processed it in luminance. Pretty good data considering light pollution. https://www.flickr.com/gp/rkonrad/m1662A.

How did you achieve the colour balance?

Richard

[almcl]

Thank you, Richard!

I think (from that very helpful video you posted) that you use a star mask to set the white point in the color module?

I didn't do that on this one but just went straight into the color after the Life module had knocked back the stars a bit. It reduced the red channel rather too much to my eye, so so I reduced the slider slightly and increased the saturation. I then found the black point was a bit too black, so gamma corrected this in the develop module. I still think it's a bit washed out and with more care could probably be improved.

The problem I have with this target is that the light pollution filter that I have to use does distort the colour balance significantly (and this may be so for the OP as well) and so using the stars to set a white balance may not work quite so well.

Be interesting to get Ivo's views on this, too.

[Burly]

In DSS settings >cosmetic tab have you got hot pixel checked ,try unchecking both hot and cold pixels then restack i see halos around back ground stars and for mre this was the culprit

[Andre27]

Good Day All

Thank you all for the response and advice , I shall re-stack again and perhaps not combine the two different exposure stacks but process them separately and then combine should the shorter exposure image add any value to the image.

After many , many hours of tinkering with this image this what I could get from the data. It is most certainly not on par with other DSLR images I have seen with far less integration time , but I think the high temp of the image frames and the light pollution is the Achilles Heal of this image stack.
I had the same issue with Rosette Nebula which is in close proximity.

I think I am going to overlay Ver 1 over Ver 2 and replace the stars as far as possible and be done with this image should the re stacking and reprocessing not deliver better results.

Thank you again for the assistance

Andre

Ver1.jpg (98.13 KiB) Viewed 8302 times

Ver2.jpg (95.97 KiB) Viewed 8302 times

[Rkonrad]

Hi Andre,

I'm not sure which programme you used to stack but if you used DSS it's important not to select "white balance". Not sure if you responded to my previous question on that.

Cheers

Richard

[admin]

I use a UHC filter that has proven to be very effective in my other images , I have however had a similar issue with the Rosette Nebula which lays in the same direction as the Wide Field Horsehead/Flame nebula which is the image in discussion.

First off, color balancing data that has been acquired through a UHC filter will be fundamentally impossible to color balance - parts of the spectrum are missing, for example;

That said, in most cases decent hints of color (pertaining to different star temperatures, reflection and emission types) can still be recovered.

It appears your dataset has indeed been filtered by some filter. Below is the default color balance the Color module came up with (for Restacked version);

Combined Flame 300 And 480 13hours N0 drizzle.jpg (244.3 KiB) Viewed 8290 times

It shows red H-alpha emissions, and the orange glow of the flame nebula. Blue reflection nebulosity is visible and some star color (mainly just orange and blue) is visible.
To be clear, I did not touch the color balance (it is honestly rarely needed).

300 & 480 Sec exposures

Please don't mix different exposure lengths. It has an adverse effect on noise propagation and star shapes.

The image above was very lightly processed, just to get to the Color module.
--- Bin
Parameter [Scale] set to [(scale/noise reduction 35.38%)/(798.89%)/(+3.00 bits)]
Image size is 1517 x 1010
--- Crop
Parameter [X1] set to [108 pixels]
Parameter [Y1] set to [86 pixels]
Parameter [X2] set to [1490 pixels (-27)]
Parameter [Y2] set to [959 pixels (-51)]
Image size is 1382 x 873
--- Wipe
Default
Parameter [Dark Anomaly Filter] set to [6 pixels]
--- Auto Develop
Parameter [Ignore Fine Detail <] set to [1.9 pixels]
Parameter [RoI X1] set to [236 pixels]
Parameter [RoI Y1] set to [240 pixels]
Parameter [RoI X2] set to [925 pixels (-457)]
Parameter [RoI Y2] set to [699 pixels (-174)]
--- Color
Parameter [Dark Saturation] set to [6.90] to introduce more color into the darker regions.
Parameter [Saturation Amount] set to [248 %]
Parameter [Blue Bias Reduce] set to [1.03]
Parameter [Red Bias Reduce] set to [4.89]
--- Wavelet De-Noise
Default
Parameter [Grain Size] set to [7.1 pixels]

Hope this helps at all...

[Andre27]

Thank you Ivo

I am going to re stack with settings suggested by the participating forum and give it another go.

Yesterday I had a meeting with the municipality and I am hopeful that the three streetlights closes to my property are going to be replaced with shorter poles and light heads that shine down instead off lighting up my backyard and all the trees in the vicinity.

Perhaps then I should be able to use a LP filter that allows for easier editing of my data and is less restrictive as far as color manipulation is concerned.

Thank you for your assistance , it is appreciated.