Xiga

Thanks Steve! 🙂

Thanks Dave! It must be whatever I did with the dust lanes that gives it the impression of moving, lol. I suppose it does have a certain 3D feel to it. As for colour, I've always found the D5300 very easy to deal with in this regard. Because it effectively has an ISO-less sensor, it means I can shoot everything at ISO 200, which gives me loads of Dynamic Range to work with, but still with very low read noise. It really is a brilliant camera for all it costs. Even though I have a (new to me) Qhy163c sitting in a box, still unused, I somehow doubt I will ever love it the way I do the wee Nikon 😀

Thank you Adam. I have to admit it did come out a lot better than I was expecting myself. Being a PS user, my workflow can be a bit all over the place tbh, lots of micro-decisions being made constantly, so hard to nail it down exactly, but I will try and post up a rough guide on the Image Processing forum over the weekend.

No need to apologize. It gave me a chuckle 😁

It's Ciaran, actually. But you can call me Steve if you want, I won't mind. Especially if it means i get to inherit his street cred too 😅

Thanks Adrian! Wow, thanks Rodd! I think some of it is down to the orientation. It was only at the very end of the workflow that i decided to flip it, and the impact was immediate. Straight away it just looked way better! So i just went with it. Thanks Richard. So it sounds like i may have, by chance, just stumbled across a similar workflow to yourself. This is most pleasing, considering the standard of your work! As you say, Starnet is truly amazing. Up until this image, i had only ever considered using it on NB images. Not any more. It's going to get used a lot from now on! 🙂

I think you've done well here Dave. It may be faint, but the dark, dusty regions are still showing nicely. This is just a very challenging target from UK skies. The only answer, is either lots of aperture, or lots of exposure! I made the mistake recently of underestimating it. On my one and only night of imaging from the last 6 months, i tried my hand at it. Managed just shy of 4 hrs (but due to technical issues, half of that wasn't really useable). In the end i couldn't even make an image i thought was worthy of posting. I'll know next time to stick to the brighter targets!

As Steve says, you're definitely being too harsh on yourself Rodd. The detail in your version is actually remarkable, and i like how you've managed the brightness levels too. I certainly won't show you my DSLR attempt having see yours! (there's always a bigger fish, isn't there, lol)

Amazing. Most people say not to even bother trying to image with a C8, but your images show that it's capable of so much in the right hands. As a fellow C8 owner, i'm definitely feeling some 'sky & mount' envy right now! lol. Great image.

Hi guys Still no chance of getting any imaging done, so what better to do than revisit old data! I thought I'd go back and re-do M31, as it is my longest data set (by quite some distance) for a true colour target. I usually only end up with a couple of hrs to work with, but on this occasion I was able to get 7 hrs worth, all on a single night with no moon about. Plus, I was never truly happy with how the original came out, but I kept putting off a re-process as I remember finding M31 an absolute beast to process at the time! 24 x 300s with an IDAS-D1 (ISO 200) 51 x 360s with an IDAS-D1 (ISO 200) Nikon D5300a SW 80ED, w FF/FR HEQ5-Pro Captured with Sequence Generator Pro, Calibrated and Stacked in Astro Pixel Processor, and processed in PS. The main difference this time, was I was able to deal with the stars properly (this time last year I clearly hadn't a clue, lol). I also did a Bayer Drizzle stack this time, as I had so many subs to work with, and it seemed to bring out some finer scale details. I used a new processing technique which I came up with (i'll try and post a thread on it when I get a chance), but basically it involved using 2 layers - the bottom layer was a lightly arcsinh-stretched one (mainly for the stars, but it also still contains sky background and nebulosity), and the top layer was a starless one (made in Starnet++) but with the blend mode set to 'Screen'. With further Curve adjustments to both layers, it was possible to push the nebulosity but keep the small stars shining through, and without dark halos. So the image below has had no star reduction applied. More experimentation needed on this one though. Oh, I also flipped it horizontally this time (not sure if that's a faux pas!). I prefer this orientation, but then again it could just be because my brain thinks it looks more interesting by virtue of just being different. C&C welcomed (I have a tendency to push my images just over the edge if i'm being honest) so i'm always on the lookout for feedback. CS! https://astrob.in/373639/H/ ps - Here's a link to the (now clearly awful) version from last year for reference. https://astrob.in/373639/0/

The basic gist of trying to tame the stars during stretching in PS, is as follows: 1. Make a copy of the background layer. Then copy the image (CTRL-A, CTRL-C), create a Layer Mask, select the Layer Mask (ALT-Click) and paste it into it (CTRL-V). Then invert the mask (CTRL-I). You should now have something like this: Use a Levels adjustment to bring the White and Black points in, and then add a slight Gaussian Blur. Start out with something small, like 0.5 pixels. Then select the Background copy layer and apply your first stretch. I usually use Arcsinh 10 stretches myself. Flatten and repeat as many times as is needed until you're happy it's been stretched as far as you want. Remembering also that each time as the stars get a little bigger, you will need to gradually increase the pixel radius of the Gaussian Blur too. This is where some experimentation will be needed. For example, i sometimes find it's a good idea to not mask every single stretch, and let the odd one or two small stretches go unmasked. The payoff of slightly bigger stars is usually worth it in the end, as they are more likely to come out a bit more natural, and in any case you can always just use a bit of star reduction to compensate. Now, depending on how well it goes, you will almost definitely still get at least some harsh edges to the outer core of the larger and medium-sized stars. It's really hard to eliminate this entirely, so if the no. of worst offenders are not too high, you can just use the Blur tool and manually soften them up a bit. Here's an example of how this method looks compared to a normal DDP stretch. This is a 200% crop of a small area of your image. You can see that the really large star (which is over-exposed) has not fared quite so well, this is one that would need a little cosmetic touching up, but the improvement in the rest is pretty apparent to see. HTH.

Ok here's how to calibrate star colours in SiriL. 1. Start by changing the working directory. There's a 'Change dir...' button at the bottom of the main window. 2. Open your linear, unstretched RGB file (File->Open). It doesn't have to be a .ft file, it can be a .tif file that you manually created in PS by combining the R,G, and B channels, but it must not be stretched in any way. 3. Run 'Image Processing->Color Calibration->Photometric Color Calibration'. 4. Search for your target, in this case i just typed 'Iris Nebula' and it found it in the Simbad database. Select it to populate the RA and Dec co-ordinates. Enter the FL and PIxel size and hit OK. Note, if it fails, then Untick the 'Auto' checkbox and increase the Limiting Magnitude. I've never had it fail with 17 so far. That's pretty much it! You will know that it's worked by the info shown in the Control Panel (see below). The only thing left to do is then Right-Click anywhere in the 'RGB image' window and choose to save the image as either a fits or tiff file (note, it can only do 16bit, but that shouldn't be a problem if you intend to move into PS from here on anyway).

Forgot to say, the very first thing I did on the unstretched linear RGB file, was to run it through Siril's Photometric Color Calibration routine. I much prefer it to APP's version (which is still a bit too unclear I find). Thankfully, Siril is free and the routine is really simple and quick to do. I'll try and describe that part later tonight too.

I find that DDP really only works best with NB data. Unfortunately, when it comes to Broadband data, you really have to go Full Manual and take control of the stretch yourself, otherwise the stars will bloat (even when properly exposed). I’m not really familiar with LRGB workflows tbh, but I think what I did was first stretch the Lum, but masking it along the way to keep the stars from bloating. Then I used Mark Shelly’s excellent colour-preserving arcsinh stretch on the RGB stack (which I combined manually in PS). Once I got both histograms roughly equal, I layered the Lum on as a Luminosity layer and then just proceeded with the usual saturation, contrast enhancement, NR etc. The method for protecting the stars is far from perfect though. It can be really hard to prevent them developing a hard edge on the inner-core, and sometimes a little cosmetic correction can be needed on the biggest stars. I’ll try and post a more detailed description later tonight with a few pics for you.

Ugh, too saturated now that I see it in the cold light of day. It was a joy to process though 😃

Hi Adam Without seeing a Raw sub it's hard to say for sure, but based on the number of saturated stars i'm seeing in the stacks, i would say you're definitely over-exposing at 300s. You should take some test subs and aim for an Avg ADU that gets you to somewhere between 5 and 10 x RN^2. Then stick to that for most targets. That being said, you can mitigate the big stars somewhat by masking them a bit while stretching. There's only so much you can do of course when they're over-exposed, but i had a go myself anyhow (thanks for sharing!). The key is to use an inverted, blurred mask of the image itself whilst you do an arcsinh stretch. You need to be careful though, as it will give you harsh cores to the stars if yo're not careful. In any case, it's crazy how much this data can be stretched, and with so little exposure! You've got a great setup Adam, looking forward to your future images! 🙂

Sorry Adam, bit late getting to this one! No the reason i asked wasn't really anything to do with how the Ha images looked. It was more just that i thought those using really fast optics had to use them tbh. I certainly don't think your images were lacking in contrast, even despite the fact the integration times were low. In fact, if anything i think they show that your setup is definitely working, so i don't think you've anything major to worry about. The real (hard to answer) question is, would the high-speed filters get you the same result, only quicker (not that big a deal, given how much aperture you have to hand), or would they give you more contrast. I don't know enough of the theory to answer that, and you may have a hard time finding real-world tests that others have done to compare to. I suppose you could always just get the Ha filter to start with, it is of course the most important one, as it often doubles-up as Luminance as well. And if you decide to sell the dual 80ED rig, then you could probably sell your current Ha filter as well to help recoup some of the cost. In any case, i look forward to seeing how you get on. Good luck! 🙂

Zero imaging getting done these days unfortunately. Baby no.2 arrived a couple of months ago so life's been a tad busy ever since, lol. I've stopped even checking the forecast now tbh. I still haven't actually got to use the Qhy163c I bought 2nd hand back in April 🙈 Things could be different if I had the option of imaging from home, but there's just too many trees, roof tops, LED lights etc to make it worthwhile. Looks like the hobby is on hiatus for a while.

Absolutely amazing Richard. I haven't been around much recently, so I must have missed this. Your persistence has really paid off. This is a truly excellent mosaic, as there are objects of interest to the left, right, and centre, which makes for a really well-balanced image. And of course the processing is top notch as always. ps - Just seen that this made IOTD on Astrobin on Nov 14. Congrats!

Adam, what NB filters are you using? Does the fast F-Ratio of the Epsilon not need specific NB filters? I think Baader do a special set of Fast NB filters for anything under ~F3.3. I'm not sure just how much of a difference they make, but it might be something to look into.

I agree the details are better in the HSO version, but rather than throw out the SHO version completely, what about mixing the details of the HSO version, with the colours of the SHO version (with a few tweaks and a bit more green taken out). Something like the one below? (hope you don't mind, i just combined the 2 Jpg's in PS. only spent about 10 mins). Really great image btw!

Very nice Adam. Great to see the Epsilon working so well! For those looking to use Starnet++ as a standalone package without P.I, it's actually really simple, and doesn't require any knowledge of DOS commands. Basically, after downloading and extracting the files, you should have a folder that looks like this: If this is NOT what you have, then you haven't downloaded the right package. In the example above, i copied in Ha and Oiii images, ready for star removal. They originally had very long file names, having come from APP, so i always rename them as just Ha and Oiii at this stage just for simplicity. Then you just need to Right-Click the run_mono_starnet.bat file and choose Edit. It should open in Notepad, and all you then need to do is check/edit the coding so it looks like the one below: So all you need to do is reference the exact filename of the file you are removing the stars from (Ha.tiff in this case - and be careful, make sure you get it exact, as sometimes tiff files can have 1 or 2 f's), and then immediately follow it with the name you want the starless image to have, in this case just exactly the same but appended with -s Save the batch file and close it. Now all you should have to do is double-click it to run it. Now, for future Ha images you shouldn't need to edit anything, and for Oiii and Sii, all you need to amend in the batch file are a few characters. The method works exactly the same with an RGB image, except you have to edit the run_rgb_starnet.bat file instead. HTH CS.

I have a spreadsheet where I calculated the optimum Median ADU level at each Gain and Offset preset that i have saved on the camera. That's how i knew to aim for an ADU level of 1400ish, and after the post-flip subs (which came out ok), I now know that 2 min subs are pretty much perfect for my skies. Now, what's even weirder, is that, even at the Max Gain preset I have saved (300), with 2 min subs I should only be seeing a Median ADU of ~6k, still way short of the ~10k I was getting. I wonder did the Offset get jacked somehow on it's own? At Gain 80, by my calculations it would take an Offset in the region of 300ish to get the level of ADU I was getting. I can't do anything about the Flats, but I might try shooting a few Darks with Gain 80 Offset 300 and see how they calibrate. This was my 1st light with a cooled Cmos. Are they known to be buggy from time to time?

I can't actually check this atm as I keep my gear at the old family home where I do all my imaging. But I don't think the meridian flip had anything to do with it tbh. I suspect that the camera was just misbehaving in some weird way, and for some reason, when the sequence aborted and then got restarted, it suddenly started working correctly. It just so happens that this was punctuated with a meridian flip. I don't think it was due to an external light either, although that's a good shout. The scope was setup in the normal place, in the driveway pointing South. No passing cars or lights could interfere. Just the usual streetlights, which are all shielded from direct view by roofs, and I use an IDAS-D1 anyways and it's always worked very well for me.

I too thought it must have been clouds, as there were definitely intermittent clouds about, and enough to lose the guide star shortly before the flip. But most of the time it was clear skies and no moon. I've also checked a good sample of the subs, from the beginning, middle, and end and every one shows the same weirdly boosted signal, so it's not down to clouds.