Eastern Veil in Hubble Palette

[Joel Shepherd]

I'd be grateful for constructive criticism on this. This is a Hubble palette rendering on the eastern Veil Nebula, starting from 2 hours each of Ha, S-II and O-III. Fairly straightforward processing in PixInsight, including deconvolution, DBE, stretching, PixelMath combination, and then a bit of work with a star mask applied to reduce the orangish hue in the stars. One thing I could use some suggestions on is further taming the star colors, or a better approach to handling them in general. Also, to my eye there is a decidedly blue hue on the right 1/5th of the frame, which is a bit of a mystery to me. Any comments on whatever strikes your eye would be appreciated too!

 

[wimvb]

Nice image with lots of detail and round stars. I am looking at it on my mobile phone so can't be too sure about colour. It seems you have a few green stars which is not natural. I agree on the right side being blue in background and the left side green. How did you map the colours?

Thanks for sharing. 

Wim

[Joel Shepherd]

Thanks for responding. I used pixel math to combine the channels, 20% Ha and 80% S-II for red, all Ha for Green and all O-III for blue. Before combining I did deconvolution, background extraction, stretched and did a linear fit (histogram alignment) before combining. Maybe the O-III stretch was off? I've tried to use a star mask but it seems to treat the larger stars as nebulosity so when I try to adjust the color of the stars alone the larger brightest ones aren't touched ... hence their green glow. Would a different type of mask work better?

[wimvb]

19 hours ago, Joel Shepherd said:

Thanks for responding. I used pixel math to combine the channels, 20% Ha and 80% S-II for red, all Ha for Green and all O-III for blue. Before combining I did deconvolution, background extraction, stretched and did a linear fit (histogram alignment) before combining. Maybe the O-III stretch was off? I've tried to use a star mask but it seems to treat the larger stars as nebulosity so when I try to adjust the color of the stars alone the larger brightest ones aren't touched ... hence their green glow. Would a different type of mask work better?

Sorry, can't help you here, since I only use a DSLR for imaging, so no narrow band. From what I've read in other posts, stars in narrow band images are often replaced by stars imaged with RGB filters, to get the star colour right.

As for star masks, I also find it difficult to get large stars included using the starmask tool. You could try using the RangeSelection tool. If you clone the unstretched image (or one of the channels) and stretch it using HT. Move the blackpoint such that most of the nebulosity is black, but stars remain. You can then use the clone as a starmask, or use RangeSelection on this image to create a mask that only contains the brightest (and hopefully largest) stars.

The screendump shows:

original RGB image with STF applied through HistogramTransformation. Note the uneven background which was not corrected for this example.

HT tool shows the histogram after stretch and the settings that will be applied to a clone.

Also shown is the clone that was black clipped with the setting of the HT tool. This image is the basis for creating a starmask.

Hope this helps.

[Joel Shepherd]

Ah, interesting: haven't had a chance to try that approach but I will in the next night or two. It would potentially solve the "big stars appear as luminosity" problem. I'm also wondering if using histogram stretch on the individual channels before combination was a boo-boo, or that at least an adaptive stretch -- if I can figure that tool out -- might do a better job of keeping background levels consistent (possibly helping with the blue-ish tint on the right). And wondering whether doing a linear fit was a good idea at all.

I've heard of substituting RGB stars in narrowband images, and am hoping to avoid that largely because of the additional capture time. E.g. I basically blew my allotment of imaging hours in June in collecting the 6 hours of NB that went into the Veil ... cool with one month projects at this point, but not so keen on investing more.

Thanks again; will report back with the results of trying a range mask.
 

[Joel Shepherd]

Here's a second attempt that I think has a little better star color. Had to kind of back into it though. I wasn't able to get a range mask to include the stars but not the nebula ... though I have an idea of how to do it now (make a duplicate, remove the stars, then subtract that from the original which should leave just stars behind that can then be used to make a mask). Instead, I eventually figured out how to get a decent star mask (high noise threshold, high scale, low smoothness), masked the image, applied SCNR which addressed much of the green tinting, and then did a bit of curves adjustment (still with the mask) to tone down the remaining red/orange. The blue haze on the left seemed to fade off after doing a DBE on the combined image.

Thanks again for the suggestions: even if I ended up going a different path it helps to talk through the problem.