Synthetic Lum from narrowband + RGB data

[Dimitris Platis]

I usually create Synth Luminance by the Image Integration tool adding my normal Luminance with R,G and B master FITS.
Now, would this offer an advantage (eg higher SNR) if I do the same with Ha plus R,G and B master FITS?
Would it retain the Ha sharpness and detail?
The catch is that the RGB FITS obviously do not share the extended structure that Ha has....

Would PixelMath work better by using some formula that would only add signal and not average it?
 

[RickS]

It will increase SNR but you will probably lose some structure and detail.  Try it and see if you like it    This may work better for some objects than others.  I have done it on very dim objects that don't have a lot of sharp detail and found it worthwhile.

Using a PixelMath formula will give you flexibility to control how much each filter contributes to the result but won't maximize SNR.  There's no advantage to adding rather than averaging linear data.

Cheers,
Rick.

[chrisvdberge]

I usually create Synth Luminance by the Image Integration tool adding my normal Luminance with R,G and B master FITS.
Now, would this offer an advantage (eg higher SNR) if I do the same with Ha plus R,G and B master FITS?
Would it retain the Ha sharpness and detail?
The catch is that the RGB FITS obviously do not share the extended structure that Ha has....

Would PixelMath work better by using some formula that would only add signal and not average it?
 

Depending on the target you might get a good (best?) result with using Ha as luminance. The R channel in any case won't add anything most likely if you have h-alpha as well.