NGC 281 LRGB

[discardedastro]

First one of the season! Back up and shooting on my rebuilt mount.

Setup is my usual: ASI183MM-PRO imaging, ASI174 Mini guiding via an OAG, Paracorr coma corrector, ZWO mini EFW with Baader LRGB filters. Scope is 200PDS with EQ6-R Pro mount. Focuser is a Baader Steeltrak driven by a Primaluce Sesto Senso.

Edit: Almost forgot! Acquisition was 40xL, 20xRGB, all 120s exposures, so a total of about 3 hours of integration, bortle 4 skies and a half moon.

Software is Ekos/KStars and INDI on a collection of Pis, PHD2 for guiding using the new multistar feature, and post-processing exclusively in PixInsight. Pretty simple workflow - calibrated, DBE'd, used MureDenoise, tried deconvolution but found I couldn't lose the ringing too readily from the brighter stars and so left it out. Photometric colour calibration. Combined LRGB, final DBE and a bit of curves/saturation to taste, and final denoising with MLT.

Really happy with this, not least because it's the first shot I've done on the EQ6-R I rebuilt!

Edited September 28, 2021 by discardedastro

[wimvb]

Very nice image. Imo, deconvolution on nebulae doesn't add much, other than perhaps reducing the stars a little.

Just out of curiosity, how many Pis are a collection?

[discardedastro]

Yeah - I was struggling to avoid unnatural-looking stars with deconv. I really need to spend some quality time with the Deconvolution process to work out a good base recipe for my images, heavily oversampled as they are the defaults don't work terribly well.

Just the two Pis really - one on the telescope itself running INDI and all the device drivers, one in the house running KStars and PHD2. I also have an all-sky camera box which is run by a Pi, and the whole setup relies on a big Supermicro JBOD/NAS (30T of ZFS storage, ish, for now) to store raw data once captured and a big Ryzen workstation for PixInsight. I VNC to the KStars Pi to drive the capture process, and a script shoves all the raw data and logs/configuration from all the programs and devices into the NAS nightly.

[wimvb]

9 hours ago, discardedastro said:

Just the two Pis really - one on the telescope itself running INDI and all the device drivers, one in the house running KStars and PHD2.

Similar to what I had until a few years ago. I went from running Kstars indoors to running it all on the Pi on top of my telescope, because whenever I lost wifi connection to my setup, the whole imaging session would come to a halt. So now I have the whole data capture software, including Kstars and PHD, on one Pi on the telescope and I connect to that Pi with RDP. If I lose connection, any sequence will continue running. At the end of an imaging night, or the next morning, I transfer all image files with FileZilla to my laptop.

9 hours ago, discardedastro said:

Yeah - I was struggling to avoid unnatural-looking stars with deconv. I really need to spend some quality time with the Deconvolution process to work out a good base recipe for my images, heavily oversampled as they are the defaults don't work terribly well.

If you PM me and send me the L master, I can have a look at it, and maybe come up with a recipe. But as I said earlier, deconvolution on nebulae won't always improve the image. And in your case, the diffraction halo around the bright stars complicate the process. If you don't mind tinkering, you could look into making a ring that sits on top of the mirror clamps, so that these are hidden. That way you can avoid the diffraction from the mirror clips.

 

Cheers,

Edited September 30, 2021 by wimvb

[discardedastro]

8 hours ago, wimvb said:

Similar to what I had until a few years ago. I went from running Kstars indoors to running it all on the Pi on top of my telescope, because whenever I lost wifi connection to my setup, the whole imaging session would come to a halt. So now I have the whole data capture software, including Kstars and PHD, on one Pi on the telescope and I connect to that Pi with RDP. If I lose connection, any sequence will continue running. At the end of an imaging night, or the next morning, I transfer all image files with FileZilla to my laptop.

If you PM me and send me the L master, I can have a look at it, and maybe come up with a recipe. But as I said earlier, deconvolution on nebulae won't always improve the image. And in your case, the diffraction halo around the bright stars complicate the process. If you don't mind tinkering, you could look into making a ring that sits on top of the mirror clamps, so that these are hidden. That way you can avoid the diffraction from the mirror clips.

 

Cheers,

Yeah - I'm fully cabled end-to-end with good quality Ethernet so connection dropouts aren't really a thing, but if it were a bit spotty then doing it all at the scope end would definitely be sensible!

integration_lum.xisf

L master's above - don't mind sharing that for all. I've had good results in some cases with deconvolution on nebulae but it's been fairly hit and miss in my trials.

Making a ring over the mirror clips I can definitely do pretty readily - I need to take the mirror out for a clean at some point anyway (it's been a few years and it's not had an easy life) so could sort it then - I figure a flat black acrylic ring in 2mm acrylic I can get cut very easily to the right size and then paint in ultra-matte black and secure with extra-long screws to the mirror clips or something. Would also help cut down stray light from the back of the tube, though that's also largely covered. It hadn't clicked that the mirror clips were what I was seeing there! Diffraction spikes from the spider are of course largely unavoidable, but I can fix the clip issue at least.

[wimvb]

3 hours ago, discardedastro said:

and secure with extra-long screws to the mirror clips or something

It occurred to me that if you make a ring out of plastic, you don't really need the clips anymore. You replace them with the ring. As with tht clips, the ring should not press down on the mirror, or you get so called pinched optics.

[discardedastro]

5 minutes ago, wimvb said:

It occurred to me that if you make a ring out of plastic, you don't really need the clips anymore. You replace them with the ring. As with tht clips, the ring should not press down on the mirror, or you get so called pinched optics.

That's a thought. Bit of neoprene or cork on the back so the contact surface isn't liable to scratch the mirror and damage coatings, and you'd be set. Hum...

[wimvb]

I had a look at your data, and as I suspected, the only thing that deconvolution does is make the stars a little narrower and a little brighter. Since most of the nebulosity is in the midtones, deconvolution will only sharpen the brighter details, but at the cost of increasing noise in the rest of the nebula. Deconvolution works best on galaxy images and planets, where ther are bright details that need to be enhanced.

Here is a before and after, with a soft STF applied

Since I applied the same stretch to both images, you need to look very carefully. You will see a difference in the smaller stars, the diffraction spikes near bright stars, and in the brightes details of the nebulosity. Medium sized stars will also appear a little narrower. Imo, if you can see that deconvolution has been applied to a dso image, it's overdone. The real work will be done in the non linear stage with subtle enhancements of the deconolved image, like sharpening and local contrast enhancements. And in retrospect, here I overdid it a little. I had a version with 10 iterations and global dark deringing set to 0.05, which I think came out better. Here, the noise starts to show.

This is the PSF that I used (extracted from low brightness stars), and the settings of the deconvolution process.

deconvolution.xpsm

With an image like this, I probably wouldn't collect any luminance at all, but only RGB and Ha.

Edited September 30, 2021 by wimvb

[wimvb]

37 minutes ago, discardedastro said:

That's a thought. Bit of neoprene or cork on the back so the contact surface isn't liable to scratch the mirror and damage coatings, and you'd be set. Hum...

If you attach the ring to the stems that hold the clips, you tighten the screws to the point where you still can slide a thin paper between the mirror and the ring (or clips). Any tighter and you risk pinching the mirror. Theoretically, you won't need any neoprene or cork. But in practice, any ring will warp slightly over time, and it's better to be safe.