Should I use LinearFit on bicolor images?

[astrovienna]

I have Ha and OIII masters on M57 here:

https://dl.dropboxusercontent.com/u/58468743/M57HaOIII.zip

Should I use LinearFit on these?  If so, fit to which one?  I found the colors, after an HOO palette merge in LRGBCombination, to be very different depending on which one I fit to.  I don't know that one was right and the other wrong, but it would be helpful to get some insight into the theory on this.  Thanks.

Kevin

[RickS]

I often use LinearFit to roughly equalise narrowband masters before combining them to produce an initial colour combine.  I don't know that there's any theory that will tell you which to use as the master.  I normally pick the one with the highest mean value but not for any particular reason.  I typically do significant colour adjustment afterwards so I'm not looking for the combine to be my final result.

Cheers,
Rick.

[vicent_peris]

Hi Kevin,

LinearFit should never replace the use of BackgroundNeutralization and ColorCalibration. As Rick says, using LinearFit you simply have a rough equalization of the RGB channels. This way, you don't have any precise background neutralization nor white referencing on your image. The way to go with this image is to create a palette with each wave band (with PixelMath or ChannelCombination) and then applying BackgroundNeutralization and ColorCalibration in the usual way. I attach a result with your image, placing the H-alpha emission in the red channel and the O-III in the green and blue, and assigning the white reference to the nebula.


Best regards,
Vicent.

[astrovienna]

Thanks, Rick and Vincent. That's very helpful.  Vincent, I have another question about how you stretched the data in your image, but I'll make a new thread on that.  That question is a LOT harder!

Edit:  Vincent, just to clarify, is LinearFit just an easy but inaccurate method of color calibration?  I find BN and CC pretty easy, so I always do them.  Does doing LinearFit before the color merge/BN/CC add anything?

Kevin