How to create a false luminance in photoshop from narrowband images

[oldannie]

If I want to create a false luminance channel from grayscale RGB images I usually go to channels , click on the dropdown menu, choose 'merge channels' change 'multi channels' to RGB, map the RGB images to the correct channels, ie green to green etc and everything works fine.  Am trying to process my first narrowband image and would like to create a false luminance with my stacked Ha, OIII & SII subs but when I go to 'merge channels' it won't let me move from multi channels and change the number of channels  from 2 to 3!  Have obviously not understood something about narrowband processing and would be grateful for some help!

 

Annie

[vlaiv]

Here are couple of approaches that should work, but I'm not sure how to do it in PS.

First is very simple and it should work very good, but check your data first if this is an option - Ha is usually the strongest signal out there - you can use it as luminance layer - so no additional tweaking necessary. Just make sure that there is Ha signal in all the places in the image where there is both OIII and HII. If Ha covers whole image - since Ha is the strongest - it will have best SNR.

If above is not option for you (data will not allow for it), consider doing following:

Create mono image, paste each image as layer (so three layers in total) - set first to normal blend mode and two above to addition mode. Do this prior to stretch and flatten image - then stretch.

This method will give you sum of Ha + OIII + SII layers. You can also use opacity sliders for top two if you want "weighted" sum (you might decide that one of channels adds very little in terms of signal but adds a lot of noise - then you want to include it with smaller weight).

Third possibility would be to combine RGB image from channels, and then do LAB decompose on that combined image and use L component of LAB as your luminance layer. This is similar to above, but you don't control weights - it is controlled by RGB->LAB conversion where not all colors contribute to luminance in the same way, but rather as we would perceive them. This works good for RGB images, but for false color, you want your best channel to be mapped to green - as eye is most sensitive in green. Standard HST palette does this - it maps Ha to G (SHO / RGB, so H ends up being G channel), and green contributes the most to luminance (eyes are most sensitive in green wavelengths) and Ha is usually channel with best SNR.

[oldannie]

Some good information Novi - thanks a lot.  Will work through what you suggest.

 

Annie

 

 

[MartinB]

The usual reason for not being able to merge the 3 channels to make an rgb is that one of the images is a different size or isn't greyscale.