M97

[ManuelJ]

Hi,

This is another archived image in the queue for processing. After processsing it, It feels quite different to any M97 image I have seen, the 3nm filter it's doing a good job here!:

http://www.manuelj.com/Astronomy/PlanetaryGallery/21825417_Hncww8

Regards,
Manuel.

[Nocturnal]

Very nice but I wonder what that would look like if you just used Ha and OIII on two channels rather than mapping OIII to G+B and Ha to R and L. I can see using OIII for G and B but why modify the OIII with the Ha? I mean that's what you do by making it L. OIII gets reduced by dim Ha and amplified by bright Ha, exactly the opposite of what you'd want. I know this sort of stuff is common in NB imaging but I don't understand the reasoning behind it.

25 Hours sure took a lot of commitment

[ManuelJ]

Very nice but I wonder what that would look like if you just used Ha and OIII on two channels rather than mapping OIII to G+B and Ha to R and L. I can see using OIII for G and B but why modify the OIII with the Ha? I mean that's what you do by making it L. OIII gets reduced by dim Ha and amplified by bright Ha, exactly the opposite of what you'd want. I know this sort of stuff is common in NB imaging but I don't understand the reasoning behind it.

25 Hours sure took a lot of commitment

Hi!,

I have used the OIII as L because it has much more SNR and structures, and the Ha is not adding much detail (not many different structures). By the way, I have tried to combine as you said, and there is no difference, just poorer SNR.

Thanks!,
Manuel.

[Nocturnal]

Hi Manuel,

I think this is a topic worth exploring a bit more. Maybe I don't understand what it means to make a monochrome image (your Ha in this case) the L of a color image. I think of applying L as a normalized multiplication not unlike flattening. There is probably something more complicated going on related to color spaces but ultimately bright areas in the L must brighten colors and dim areas must dim colors. I hope Juan can respond and clear this up.

[Juan Conejero]

Saying this is somewhat risky without looking at the data, but in my opinion, using one of the narrowband images as luminance (OIII in this case) doesn't seem a good idea. Adding more luminance always weakens the chrominance, and since it is redundant data in this case (that is, the same data already used for two of the chrominance components), it has no real benefit in terms of noise reduction, either. Better control should be possible with histogram and curves transformations (L curve to control lightness and c* or S curves for color saturation) and a suitable noise reduction.