NGC6188 and NGC6164 in HaLRGB

[mountainjoo (Jerome)]

This ended up being about 11 hours of integration time. I struggled to match star profiles when combining the luminance and RGB images, so ended up masking the stars before combining. I'd really like to hear any suggestions that people have for approaching this step.

Please find the full res here, any feedback would be greatly appreciated. I'd like to do a narrowband only attempt in the future.

[Atmos (Colin)]

It looks pretty good Jerome, I do agree that matching star profiles can be difficult. Have to come back to this with a bigger screen than my iPhone

[glend (Glen)]

Nice Jerome, as you know i have been working on NGC 6188 as well (it is in my Astrobin gallery top row right side in mono Ha). To me your image seems a bit dark and i think it would benefit from using Ha as a luminance layer. That way you might be able to catch more of the light streaming out from behind the cloud wall. Just a suggestion, but it is still very nice. Whoops, on the other hand, did you use Ha for luminance or a luminance filter?

[mountainjoo (Jerome)]

Originally Posted by Atmos

It looks pretty good Jerome, I do agree that matching star profiles can be difficult. Have to come back to this with a bigger screen than my iPhone

Thanks Colin. After several failed attempts at trying to match the star profiles by shrinking the stars in the RGB image or growing the stars in the luminance image, I gave up and just masked them off. Seemed to work well.

Originally Posted by glend

Nice Jerome, as you know i have been working on NGC 6188 as well (it is in my Astrobin gallery top row right side in mono Ha). To me your image seems a bit dark and i think it would benefit from using Ha as a luminance layer. That way you might be able to catch more of the light streaming out from behind the cloud wall. Just a suggestion, but it is still very nice. Whoops, on the other hand, did you use Ha for luminance or a luminance filter?

Thanks for your suggestion Glen. I haven't used the Ha channel by itself for luminance. I have taken luminance data using a luminance filter and blended the Ha into it. I imagine that I could get away in this image with just using Ha for luminance (there's just the one small reflection nebula) and masking off stars when combining with the RGB to avoid the discontinuities (which I ended up doing anyway). I'll have to try this at some point.

[gregbradley]

Nice image Jerome. I like the framing, creative and it fits in the main features.

The stars do look a bit muted. Not sure what you mean about matching the star profiles. Luminance is always sharper than the rgb stars. What you don't want is one RGB with much fatter stars than the others as that give nasty halos, like a red ring or blue ring.

In that case either reshoot the offending RGB with close attention to focus and matching the seeing as well as you can. Or doing a bit of deconvolution on the offending larger starred R G or B.

Otherwise the luminance tends to shrink the stars anyway.

The shockwave around NGC6164 is mostly in O111 and somewhat in Ha. But it takes a fair bit of exposure to bring it out. Best to use 2x2 binning on the O111 and Ha. 6188 also looks very good in narrowband. In some ways better than in LRGB.

I have just spent a LONG time on NGC6188 so these nuances are fresh in my mind. I need to collect more S11 and will have a complete set of data. 11 hours is good but narrowband is going to require another 11 hours or so.

Greg.

[mountainjoo (Jerome)]

Originally Posted by gregbradley

Nice image Jerome. I like the framing, creative and it fits in the main features.

The stars do look a bit muted. Not sure what you mean about matching the star profiles. Luminance is always sharper than the rgb stars. What you don't want is one RGB with much fatter stars than the others as that give nasty halos, like a red ring or blue ring.

In that case either reshoot the offending RGB with close attention to focus and matching the seeing as well as you can. Or doing a bit of deconvolution on the offending larger starred R G or B.

Otherwise the luminance tends to shrink the stars anyway.

The shockwave around NGC6164 is mostly in O111 and somewhat in Ha. But it takes a fair bit of exposure to bring it out. Best to use 2x2 binning on the O111 and Ha. 6188 also looks very good in narrowband. In some ways better than in LRGB.

I have just spent a LONG time on NGC6188 so these nuances are fresh in my mind. I need to collect more S11 and will have a complete set of data. 11 hours is good but narrowband is going to require another 11 hours or so.

Greg.

Thanks Greg. I was unsure at first with the framing so I'm glad to hear that you like it.

The muted stars is maybe a consequence of going overboard with the erosion... I've attached an image illustrating what occurs when I do the LRGB combine. The blue stars appear most impacted and seem to acquire a dark ring. Is this due to them being bloated? The blue channel had the lowest FWHM out of the 3 colour channels. Or is it just that I'm crazy and it's not as severe as I think?

I'm looking forward to seeing your rendition. I'll need to get myself OIII and SII filters before I can even consider getting the next 11 hours! Even just adding OIII to the RGB would really help with the bringing out the shockwaves in NGC6164.

[gregbradley]

I found also I got some dark rings around stars. I think you'll find that comes from the Ha. I take it this image has Ha? Do a star mask on your Ha layer and the dark rings will probably go away or reduce. Anyway that worked for me.

With narrowband you have to protect the stars from narrowband data which is a star wrecker. Mask the stars on each narrowband filter added to LRGB data to preserve the LRGB stars. Make sure the star selection size is a tad larger than the stars themselves to prevent the rings.
The erosion filter is like the minimum filter and I think its quite a destructive filter and to be used lightly and preferably not on stars at all.

PixInsight has a median selection in the Morphological transformation tool.
If you set it to circles as the structure it can improve things but it can also wreck the rest of the image. So again I suppose a mask for the stars is needed (which is not a simple task in PI and simple to do in Photoshop).

Greg.

[RickS (Rick)]

That's a nice image, Jerome. I've never had to match RGB stars to the Luminance. It's not normally that critical unless you have one colour with a FWHM significantly worse that the others and you get coloured rings around the stars.

In the small example you posted it doesn't look like an issue with the stars to me - the black point looks to be quite different. The image on the right is darker and also more detailed.

Cheers,
Rick.

[Slawomir (Suavi)]

Very nice image Jerome, I think it is good as is

I must admit though, that I am colour junkie and prefer more vivid full narrowband images of nebulae

[mountainjoo (Jerome)]

Originally Posted by gregbradley

I found also I got some dark rings around stars. I think you'll find that comes from the Ha. I take it this image has Ha? Do a star mask on your Ha layer and the dark rings will probably go away or reduce. Anyway that worked for me.

With narrowband you have to protect the stars from narrowband data which is a star wrecker. Mask the stars on each narrowband filter added to LRGB data to preserve the LRGB stars. Make sure the star selection size is a tad larger than the stars themselves to prevent the rings.
The erosion filter is like the minimum filter and I think its quite a destructive filter and to be used lightly and preferably not on stars at all.

PixInsight has a median selection in the Morphological transformation tool.
If you set it to circles as the structure it can improve things but it can also wreck the rest of the image. So again I suppose a mask for the stars is needed (which is not a simple task in PI and simple to do in Photoshop).

Greg.

The image has Ha blended into both luminance and red channels. However, before blending, the stars are essentially removed from the Ha channel so I don't think that this is the cause. Just to make sure, I tried masking off the stars in the luminance channel before adding the Ha and this still leads to similar results as shown in the comparison image I posted earlier.

I've been using the minimum filter in Morphological Transformation with circle as the structure for star reduction (in conjunction with a mask that only exposes the stars). Would the median filter be a better choice for this? I am usually quite careful with this step, but in an effort to make the nebulosity stand out I may have pushed it too far.

Originally Posted by RickS

That's a nice image, Jerome. I've never had to match RGB stars to the Luminance. It's not normally that critical unless you have one colour with a FWHM significantly worse that the others and you get coloured rings around the stars.

In the small example you posted it doesn't look like an issue with the stars to me - the black point looks to be quite different. The image on the right is darker and also more detailed.

Cheers,
Rick.

Thanks Rick. Perhaps it's the sharpness in star fall off when adding luminance that's throwing me off. The FWHM of the R, G and B channels are within 10% of the average.

Originally Posted by Slawomir

Very nice image Jerome, I think it is good as is

I must admit though, that I am colour junkie and prefer more vivid full narrowband images of nebulae

Thanks Suavi. I must be a colour junkie too - I have a craving for a couple Astrodon narrowband filters to complete the set, and a WSG-8 upgrade...

[Retrograde (Pete)]

Impressive image Jerome. Nice to see this object in HaLRGB

[gregbradley]

I've been using the minimum filter in Morphological Transformation with circle as the structure for star reduction (in conjunction with a mask that only exposes the stars). Would the median filter be a better choice for this? I am usually quite careful with this step, but in an effort to make the nebulosity stand out I may have pushed it too far.

Got it about the Luminance and Ha blends with red as well. The Ha as luminance can do it as well. I haven't really explored Ha blended with luminance and also with red. I should. Other ways are more flexible but perhaps a different result. Its worth exploring all the different ways and decide which approach you like the best.

I've had mixed results with this tool. At one point I thought I had it under control. But yes that is how I have used it. What I do notice is that whilst it can make an improvement to the star shapes it often damaged the main image. So a star mask would seem vital. Reducing the settings minimised that but not entirely. So becoming adept at doing a PI Star mask would seem to be important.

[strongmanmike (Michael)]

Well Jerome, I recon this is a fabulous image, looks pretty damn good to me. I also like the framing, really nice.

Great job

Mike

[mountainjoo (Jerome)]

Originally Posted by Retrograde

Impressive image Jerome. Nice to see this object in HaLRGB

Originally Posted by strongmanmike

Well Jerome, I recon this is a fabulous image, looks pretty damn good to me. I also like the framing, really nice.

Great job

Mike

Thanks Pete and Mike! Much appreciated.

Originally Posted by gregbradley

I've been using the minimum filter in Morphological Transformation with circle as the structure for star reduction (in conjunction with a mask that only exposes the stars). Would the median filter be a better choice for this? I am usually quite careful with this step, but in an effort to make the nebulosity stand out I may have pushed it too far.

Got it about the Luminance and Ha blends with red as well. The Ha as luminance can do it as well. I haven't really explored Ha blended with luminance and also with red. I should. Other ways are more flexible but perhaps a different result. Its worth exploring all the different ways and decide which approach you like the best.

I've had mixed results with this tool. At one point I thought I had it under control. But yes that is how I have used it. What I do notice is that whilst it can make an improvement to the star shapes it often damaged the main image. So a star mask would seem vital. Reducing the settings minimised that but not entirely. So becoming adept at doing a PI Star mask would seem to be important.

The approach to blending that I use is explained here. It's the only method I've used but has been reliable enough for me to not feel inclined to investigate other methods. I'd certainly like to hear how others are blending narrowband with broadband data though.