Struggling with color calibration

[pbkwee]

I have a 1.6MB binned rgb image at http://meridianflip.com/raw/ngc5128-r10x300s-g10x400-b6x400s-small.tif  It is the R G and B integrated images.  No stretching or other processing. 

I can't seem to get the color balance correct.

I tried dynamic background extraction, color calibration, assisted color calibration and channel match.

I have more red and green than blue frames.  And the blue frames are perhaps affected by a rising sun and foggy morning.

Any step by step advice for me, a new PI user, on the workflow to get the right color calibration?  Or are these image sets not salvable?

Any general advise?  e.g. if I take different length exposures for each channel, is it possible to wrangle a correct color balance from them in PI?  Or do the duration lengths need to be correct for each channel?

[georg.viehoever]

Usually doing backgroundcorrection first and colorcalibration Second does much of the work for me. You may have to Play with their Parameters so they Identivy true background and non-saturated Stars correctly.

[pbkwee]

I manually calculated some ratios for RGB based on a part of the image I picked would look OK as white.  I ended up with this: http://astrobin.com/11652/ 

I'm not able to figure out how to use the dbe or colorcalibration to do this automatically.  If you were able to use the linked tif file and let me know the settings you used in order for pixinsight to figure out the correct color calibration that would help me understand those tools better.

[georg.viehoever]

Did you have a look at Harry's videos, for instance this one http://www.harrysastroshed.com/CC.html ?
Georg

[georg.viehoever]

It is a bit difficult to get an equal background when you only have such a small snippet, so my try at DBE was just quick and dirty with just 11 samples (the original has a strong gradient from top left to bottom right). Then did a BackgroundNeutralization with Preview1 as reference, then a ColorCalibration. You can see the settings in the screenshot. Exept for lowering the upper limits for background everything was at defaults. I find it quite okay, except for the slight ring around the core which is probably a result of the less then perfect DBE.
Georg

[Geoff]

One way that seems to work quite well if you are confident about your flat fielding is to simply adjust the RGB histograms separately using histogram stretch so that the three peaks all line up. Of course, you are then committed to nonlinear images.
Geoff