This is worth a new post! Jase very kindly e-mailed me this evening, offering to have a go at processing the Messier 16 LRGB data I acquired from suburban Brisbane last night, 23rd August. Needles to say, I eagerly accepted, anxious to see what his masterful processing skills could produce; I submitted a sows ear and I was hoping that I might get at least a cotton purse, but wasn’t expecting a silk one let alone a golden one!
I provided Jase with the already reduced LRGB fits files (Darks, Flats, Aligned, Combined). The RGB files were acquired at low altitude in the West, with hazy cloud causing some auto guiding problems, and dipping into the light pollution dome of Brisbane. So, Jase didn’t have the original data to work with, just the stacked files I e-mailed to him. Also, I submitted RGB 1:1:1 as opposed to the SBIG recommended 1:1:1.6 ratios.
Well, well, well. My goodness me – what a difference!
Look at the subtle colours in those stars and the detail, smoothness and clarity of the processing. I guess it must have been a breeze for Jase having such small lumps of 765x510 pixel data to play with compared to his most recent, vast cosmic scapes. Thanks for this Jase, you have boosted my confidence no end but sadly, have doomed me to walk the path of LRGB imaging now; no one shot colour here after seeing your magic recipes at work!
Thanks also for sending back to me, the “working” psd file including layers – I’ll spend some time with this to learn how it is done.
Cheers
Dennis
PS – I have included my original to show the comparison.
PPS – Anyone wanna buy time on a new rent-a-scope in Brisbane! LOL!
Not wanting to sound disrespectful Dennis, but my word, what a difference! Jase has managed to make this object look so much more 3D with his processing! The detail which is there in both images just wants to be noticed.
Excellent work on the capture, and nice processing Jase
Seeing the results that Jase has achieved has been quite an eye opener for me, nothing short of an epiphany; now I know how Paul must have felt on the road to Damascus – I’ve been converted to LRGB imaging hereon.
Although I have an excellent rig (EM200, Mewlon 180, SBIG ST7E) I’ve been really struggling with DSO’s at fl’s of over 1500mm and just put it down to poor data capture. Jase’s masterful processing of my M16 has now provided me with a quantum leap in understanding – it is the post capture processing that I need to focus on!
Although the RGB data I provided to Jase looked pretty ordinary to my eyes, he has used it and turned it around to produce such a stunning image compared to my feeble efforts. Jase’s act of generosity has been like having a CS3 tutor and in my case, has opened up a whole new world.
Thanks again Jase, I am humbled by your example – it is a rare gift and talent indeed, to be able to reach down and help beginners and advanced imagers alike.
A pleasure Dennis. Rarely do I gain access to data that was simple to process. Usually, I'm wrestling with gradients, blooms and noise which plainly compound ones efforts, but I enjoy the challenge. For the long FL you're running (~1800mm), I'd say you've got the rig tuned. Keep doing what you do with regards to the data acquisition - its on the money. The ST7 delivering (1.07 arcsec/pixel) provides good sampling. I wouldn't bother binning the chip at all - even for RGB data collection. You've got the sensitivity and the chip is small already.
Despite the additional work, mono chip with filters will provide the greatest level of control and is certainly my preference. Once you've been there, its difficult to go back. If anything, I'd invest into a bigger chip. Perhaps an ST10. I'd then consider buying some rental scope time from you. Sure you can get off the SBIG boat, but there is a lot to be said about simplicity of self guiding on the optical axis...certainly as you go up in FL anyway. If you have problems finding guidestars through a Ha filter, simply bin the guide chip to boost the sensitivity. Guide star selection is over rated. Just takes planning, thats all.
For the benefit of others, the processing flow was as follows (summarised). Luminance passed through two iterations of deconvolution and saved as TIFF ready for PS. RGB was colour combined and saved as TIFF. I don't log stretch any of the data. I prefer total control over data stretching in PS - the only exception to this was the follow; A second RGB colour combine was created and DDP stretched heavily with saturation boosted to 180% - then saved as TIFF. All TIFFs opened in PS. Work on RGB, levels and curves with minor boost in saturation. I carefully watched the histogram while stretch. I did not stretch the data all the way to it full potential. I then worked on the luminance and started to stretch it. The luminance was added as a luminosity blend layered on the RGB previously stretched. The luminance was duplicated so that two luminance layers existed in the layer stack. The duplicated layer was then stretched further. All non-linear stretches (curves) were performed with specific top end adjustments to manage the stellar profiles. The two luminance layers contrast each other providing depth - simply use of opacity alters the integration into the layer stack. An alternative way of providing a contrast boost is to duplicate the luminance - stretch it hard - then use the multiply blend mode. The Luminance washes much of the colour in LRGB blends. You can use many methods to circumvent this by introducing the luminance at only 50% and flattening, to greater a so called super RGB, boost its saturation, then add the luminance again at 100%. So you end up with an LLRGB image. In this instance. I took the RGB DDP stretched TIFF and boosted the saturation, blending in as soft light into the layer stack (at top). Then altered the opacity to bring back a more natural feel and lightening the image (soft light is a darkening function). I performed some minor colour balance tweaks and selective sharpening through other layers, but nothing really out of the norm. Having a look at the image again this morning. I think think the blue cast over the image isn't too attractive. Increasing the blue channel black point or dropping it using colour balance shadows would bring out the vibrant reds of the region while not compromising the subtle star colours. Personal preference anyway. There is no end of ways to process LRGB images. What is important is to keep an eye on what you are doing with curves. As you boost the luminance, you need to counteract with colour saturation. You'll find if you stretch the luminance to its full potential, you'll struggle to bring back the colour. Work methodically on both luminance and RGB data sets. Finally, don't set your black point too high during the data stretches, just enough to see what you're working with. As you come closed to the end of the routine, then raise it to remove the "whitespace". You'll find that as its raised, the richer the colours will be become (watch out for white clipping). If they become too rich, simply desaturate the image to your liking. In this entire process, make sure you're still having fun. If processing is giving you a headache and stressing you out - take a break - its not worth it.
Thank you all for your kind comments. I appreciate your acknowledgement. I'm sure I'll bring back a few tricks with me from AIC2008 to share. Look forward to seeing more of your LRGB efforts with some new found knowledge Dennis.
I have always found Jases criticisms and comments most helpful.
Yeah – the guy's a genius!
Quote:
Originally Posted by avandonk
Bit of a relief that your imaging is fine hey Dennis?
Bert
That's become much clearer now Bert! For some misguided, deviant reason, I had been expecting my raw data captures to look almost “finished” before I had even begun the processing cycle, so I have just not been too happy with them over the last couple of years, pending fine tuning of my capture technique as I gradually ironed out niggling bugs and sloppy technique.
I have made enormous inroads into getting my rig astro photography ready, but paid scant attention to image processing as the data never quite seemed up to scratch. Noise, “thin” images, hot/cold pixels, vignetting, trailed stars, etc. etc.
Although there is always room for improvement in the quality of data capture, the piece of the jig saw puzzle that was really missing for me was that I (mostly) already had the data, but I simply didn’t believe it, or better still, know it!
Jase’s process of the M16 data finally made the penny drop, so all I have to do now is to better understand the nature of the data that I am able to acquire and how to best process it, to tease out the latent “stuff” embedded in there. Shouldn’t take me longer than a few minutes of practice eh?
As you probably well know Dennis I was never happy with my widefields. I think I invented my method for HDR processing and it works very well. I am still learning and without Jases input I could have been stuck in a blind corner. His analysis by simply looking at my colour histograms was quite enlightening.
My best friend is a chef, this is like the difference between his and my cooking - masterful guys.
It raises my spirits too. Once you've tuned your gear to a certain spot - it tells me skill in processing can really improve your images discernibly. Well done indeed!
How much per hour dennis??? 
Sure you can get off the SBIG boat, but there is a lot to be said about simplicity of self guiding on the optical axis...certainly as you go up in FL anyway. If you have problems finding guidestars through a Ha filter, simply bin the guide chip to boost the sensitivity. Guide star selection is over rated. Just takes planning, thats all.
