Another different take on Ara. The weekend before lockdown I headed to Andy01's place to help diagnose some equipment issues so I took my SVX80 setup and shot some HA and Oiii NB data with the ASI2600MC (Because why not)
Andy suggested this orientation and the following morning ran through a process which gave a pretty snappy bicolour, but I wanted RGB stars. Last night was the one clear night I could see so I shot OSC data until the moon was about to rise then flipped over to Oiii for the rest of the night, as I had a limited amount from Andy's place so it was noisy. What was interesting was that while the HA looked really good, the Oiii shot at Andy's place looked a little weak and the Oiii shot here under my darker skies was so much better that I didn't include the older data in the integration. I suppose it just shows that dark sky makes a measurable difference even to NB imaging.
This is the result from 47 X OSC subs of 300 seconds, calibrated in APP and split to seperate red-green-blue files, and 29 X 600 second HA and Oiii subs, debayered in APP using the HA and Oiii algos respectively and saved as mono subs, with the whole lot integrated in a single process to give R-G-B-HA-Oiii integrations that were registered together. The R-G-B was put together as an RGB image in APP then the lot layered up in Photoshop to produce an RGB-HA-Oiii image. More data would be nice (As always) but I am pretty happy with that as my first proper crack at an RGB-HA-Oiii image.
Yes, a trifly noisy, I think the noise is sort of inescapable when shooting NB with an OSC camera (At least without a bazillion subs) with all the interpolation that has to go on to produce the data for four pixels out of one or two pixels per RGGB group of four. The straight RBG component is still a trifle noisy but it a lot smoother than the Oiii in particular, the RGB was shot in half of one night where I usually like to shoot a couple of nights on a target.
I have heard that people who bought ASI2600 mono cams direct from ZWO have started getting "Your order has shipped" emails, hopefully mine (Bought locally) is in the same batch so I can reshoot this target (At least the NB) to do a straight comparison between interpolated OSC-NB and mono NB for sharpness and noise.
Wow a lot of effort went into the processing of this. Mind blowing. I don’t have that patience
Great final result!!
There are a few steps in it, but compared to a normal mono workflow, it is really not that many extras. Really the only extra work compared to mono is saving the calibrated subs to extract the red, green, blue, HA, Oiii respectively. From there is is about the same process as anyone shooting RGB-NB would use.
Yes, a trifly noisy, I think the noise is sort of inescapable when shooting NB with an OSC camera
Not really, it's all about selective processing and layers,
Also Topaz denoise works extremely well when you remove the stars and work jst on the nebulosity. the way to process OSC is to use starnet++ work on the nebulosity and then re-add the stars later
I have never been really happy with Topaz results, though I suspect it works better with mono data than with OSC, at least with OSC I was finding that even on the lowest possible setting I was seeing all sorts of artefacts popping up.
And I think similar for Starnet++ actually, it appears from the ones I have run through it to work better on mono data than colour.
Anyway, I am not unhappy with this at this point, the processing was literally my first crack at RGB-HA-Oiii and I only had about an hour to work on it. Real life intrudes! With any luck the mono cam arrives soon and I can redo the NB subs and then put a lot more time in to the process.
Last edited by The_bluester; 07-06-2021 at 02:20 PM.
I have never been really happy with Topaz results, though I suspect it works better with mono data than with OSC, at least with OSC I was finding that even on the lowest poossible setting I was seeing all sorts of artefacts popping up. With OSC data
And I think similar for Starnet++ actually, it appears from the ones I have run through it to work better on mono data than colour.
Anyway, I am not unhappy with this at this point, the processing was literally my first crack at RGB-HA-Oiii and I only had about an hour to work on it. Real life intrudes! With any luck the mono cam arrives soon and I can redo the NB subs and then put a lot more time in to the process.
What you do is create a HA extracted image, then do a starnet++ on that, do the same with the rgb combined one,
On the Ha version do the topaz denoise, same with the rgb, then add the Ha as a luminance layer, you'll be amazed at how smooth the image becomes. As you yourself stated you only spent an hour on something new to you. It take me up to 3-4 hours to finish an image when processing. I'm very ocd with the process and will tweak here and there, even then I'm never happy. but noise can definitely be tamed. (would love the funds for the 294mm so I can then play around and use the 2 cameras together.)
I do have the Nik tools collection, I mighty try a Topaz trial again and compare the Topaz denoise with the Nik Define noise reduction. The HA and Oiii layers are starless already so not hard to give it a run on those. Job for a rainy evening and there are a few of those coming up. But I am keen to hold off and replace the NB data with true mono NB, interpolation of three pixels from one in the case of the Ha and two tto two in the Oiii surely has to have an impact, but I don't know anyone who has back to back tested to see. I am keen to be the guinea pig there.
I do have the Nik tools collection, I mighty try a Topaz trial again and compare the Topaz denoise with the Nik Define noise reduction. The HA and Oiii layers are starless already so not hard to give it a run on those. Job for a rainy evening and there are a few of those coming up. But I am keen to hold off and replace the NB data with true mono NB, interpolation of three pixels from one in the case of the Ha and two tto two in the Oiii surely has to have an impact, but I don't know anyone who has back to back tested to see. I am keen to be the guinea pig there.
I have never been really happy with Topaz results, though I suspect it works better with mono data than with OSC, at least with OSC I was finding that even on the lowest poossible setting I was seeing all sorts of artefacts popping up. With OSC data
And I think similar for Starnet++ actually, it appears from the ones I have run through it to work better on mono data than colour.
Anyway, I am not unhappy with this at this point, the processing was literally my first crack at RGB-HA-Oiii and I only had about an hour to work on it. Real life intrudes! With any luck the mono cam arrives soon and I can redo the NB subs and then put a lot more time in to the process.
If noise is an issue increasing your SNR by binning your mono channels or shooting more subs and dithering will all help greatly. Regarding noise reduction methods there are many good programs out there. Unfortunately Topaz is definitely not one of them. I would also recommend not to sharpen your nebulosity after you have removed the stars. You need the stars for proper sharpening to take place. Otherwise you'll introduce artefacts and details that are not consistent with the original starfield resolution. I see too many pictures lately with these problems. This one is not one of them.
Until my mono camera arrives (Fingers crossed it is in the batch ZWO are currently shipping) my only real option for noise reduction other than software is averaging it out with more subs, that is my preferred method actually.
That takes a while with the sky as it has been and doing 600 sec subs. If the mono cam arrives soon I might shoot another night of RGB with the OSC camera to really smooth that out, before flipping the filter wheel on to the new mono one and then spend several nights doing the NB subs from scratch. I might even have my Sii filter by then too, though it is a "Birthday" present so I am not allowed to use it until mid next month even if it does show up!
Until my mono camera arrives (Fingers crossed it is in the batch ZWO are currently shipping) my only real option for noise reduction other than software is averaging it out with more subs, that is my preferred method actually.
That takes a while with the sky as it has been and doing 600 sec subs. If the mono cam arrives soon I might shoot another night of RGB with the OSC camera to really smooth that out, before flipping the filter wheel on to the new mono one and then spend several nights doing the NB subs from scratch. I might even have my Sii filter by then too, though it is a "Birthday" present so I am not allowed to use it until mid next month even if it does show up!
Yeah it is the preferred method. There is no freebie. Shooting with an OSC and dithering makes a huge difference though because you can take advantage of bayer drizzle as well in PI when you do a drizzle integration by checking CFA. So instead of shooting 10min subs double the amount of 5min subs and do large dithers on all subs. All I'm saying is that shot has very good underlying details so it is technically very good data. Don't go butcher it with some dodgy software just on the home run.
I do dither, and reasonably heavily. A couple of pixels per sub, every sub. My first guided shot ever displayed walking noise and I trawled the net, discovered "Dither or die" and have dithered ever since (People tell me I have been inclined to dither a lot longer than that)
At the moment though it is going to be a race between the new cam arriving and the sky clearing, I really don't know which I will get sooner. I have not had the chance yet to try 5 minute subs on the NB (I used 5 minutes on the OSC data) to see what the signal levels are like and work out if shorter subs, lots of dither and drizzle processing is useful.
Probably experiment time when the mono arrives. Storage is cheap, and the drives on the actual imaging PC are 0.5tb and around half full so I am not going to fill that in a night before clearing the data.
I like this orientation of Ara, I did it myself a while ago.. it converts it from the fighting dragons to the mother bird feeding her chicks.
Nicely processed