Posted by R J Andersson at 10:42 on 2012 Aug 09
Hi Andrea,Thanks for the feedback. I agree about the artefacts – I still have work to do to refine my skills in that area. Part of the problem is deciding what the viewing distance will be and maybe I’m being too optimistic in assuming that the image will be viewed at a comfortable distance in its entirety rather than at a scale where each pixel is easily resolvable to the viewer’s eye. H-alpha wasn’t used as a luminance layer so maybe it would be of interest if I briefly described the processing. First up here’s a montage of some of the separate elements:
I don’t shoot luminance subs so I have yet to experiment with binning. I combine all my R, G and B subs to produce both a luminance image and a separate, and inevitably less deep, colour image. Apart from white balance and saturation tweaks that colour image is pretty much left alone and used as a "Color" layer in Photoshop. The luminance image for this shot was processed in PixInsight in a standard way with the extra step of creating a Star Mask to allow a spot of "HDRMultiscaleProcessing". The result was exported to Photoshop (16 bit TIFF is my preferred route) and then the stars were removed by a sequence of Dust & Scratches filter applications followed by a painstaking manual cleanup. I then subtracted the "No Stars" version from the original to produce a "Just Stars" version. The "Just Stars" version went back through PixInsight for a spot of deconvolution but I ended up applying virtually none as there weren’t any bloated stars in this image. I often pass a "No Stars" image back through PixInsight as well but this time I saw no benefit and achieved a better result by using a combination of the "Astronomy Tools" contrast enhancement "Action" and Topaz Labs "InFocus" Photoshop plug-in to boost the micro-contrast.The star removal process was key to getting the Ha image into a usable form. As you can see, subtracting the "Green divided by 15" signal from the stacked Ha image got rid of a lot of the unwanted continuum signal in the disk but I was left with quite a number of star ghosts, not shown above and presumably from red supergiants local to our own galaxy. With those removed (I used a blink comparison method by toggling a temporarily added "Just Stars" layer – see the previous paragraph) you can see the result above. There is still quite an excess of an Ha signal from the central bulge. I could have just erased it but as it shows spiral structure I decided to leave it and, fortunately, my colour choice for the Ha component didn’t upset the overall colour balance of the core when I added the Ha data as a Photoshop screen layer.The bottom right image is just a version of the originally posted image but without the Ha layer.So, returning to your comment about the lack of star colour. As mentioned above, apart from the separate H-alpha contribution the only colour in the image is due to a single "Color" layer in Photoshop. As a matter of personal preference I’m generally very unimpressed by images which show large numbers of brightly coloured stars as they look very unnatural to me. But I’ll then immediately ‘fess up and admit that my treatment of M31 is also unnatural as I’ve considerably compressed the dynamic range to display the outer disk without totally blowing the central bulge. Similarly, the micro-contrast to highlight the spiral structure is equally unnatural, not to mention the totally artificial rendition of the H-alpha intensity. I could certainly have made different choices, perhaps by applying the "Color" layer as two separate clipping masks to the "No Stars" and "Just Stars" components and changing the saturation for those two components separately. In the end I chose the simpler route and also went for a muted approach where I concentrated on getting the disk colour of M31 where I wanted it with the core hinting at yellowish and some blueish tints marking out the hot young stars in parts of the outer disk.I’m still very much a newbie at this sort of processing. So much to learn but in terms of a personal "house style" I’m getting close to where I want to be but I always value feedback such as your own to make me question my choices and, hopefully, do better in the future. :)Bob.