[RCC] I need help to improve my technique and keep learning

Let's start with the easier things: the palette is too blue, verging on the purple. The image scale is too high for the amount of time spent on it (so you might as well halve it and preserve some details), color balance is a bit off and stars both fuzzy and lacking colors (the brightest ones at least). Too much noise reduction, which you wouldn't need if you balanced the exposure to image scale, also applied where you don't need it the most. Think about adding 3 times as much exposure and possibly even more. I do hope you added the RGB data to the luminance data to boost SNR and if not then do it. And collimation doesn't look too good either.

More time and maybe resample to somewhere between 1-1.5" per pixel resolution. The more integration time you have the more you can get away with a higher resolution sampling rate, and at 6 hours youre not quite there yet. I can count ne number of nights in one hand where 1"/px is a sensible resolution with my 8" newtonian imaged from typical sea level conditions, its really quite demanding for your kit and sky.

I echo the 'fuzziness' of the stars, which indicates some issue with either the kit or the conditions. Could have been high cloud, or scope issues such as poor thermal control (lack of fan, dont know if your scope has one, if not then mount one behind the primary).

What sort of fwhm in arc seconds did this night produce? It can give a hint if there is something wrong. With my 8" a good night can be around 2.3", a decent night maybe 2.8" and bad ones can be more than 3.5" (these nights are not useful for high resolution imaging at all).

On the dark thing i will say that it is true that your camera has very low dark current but you still have to subtract offset somehow. You can do it with most stacking software but not doing it will result in failed flat frames (flats also nees this). By far the easiest and most sensible approach is to just take the dark and darkflat frames.

Hi Jordi, just to summarise and expand on some points already mentioned and add a couple more:

I recommend taking dark frames and producing a master dark. One master dark for each exposure time that you are likely to use. Save the master darks into a library and then simply pick up the correct one to use when calibrating the subs. Master calibration frames can be kept for at least three months before taking a new set.

I see that the acquisition details for M101 say that the sensor temperature was -5 not -10 as in your post. 

Are you using the Weighted Batch Preprocessing Script (WBPP) to calibrate and integrate your data? If not, please do as it is a great time saver that does all the work for you. Simply feed I all the light frames and calibration frames and press the go button.

There is no mention of flat frames in your post. Flat frames are important to correct vignetting and remove dust bunnies and can produce a cleaner master output that makes post processing easier. Flat frames should be calibrated with flat darks (of the same exposure time) for your cmos sensor. Once a library of master flats is created then simply pick up the mast flat for each new calibration run. 

There is definitely a blue cast to the M101 image and your Rosette image as well. There is no mention of colour calibration in your post. Do you use the SPCC process for colour calibration? I recommend reviewing your workflow with colour calibration in mind.

Collecting more data will help with noise management by improving the SNR. This allows for a lighter touch when applying noise reduction in post processing. Don’t forget as well that we don’t want a too smooth background as it will look false.

On the data quality issue, it might be worth while running the data through the sub frame selector process to see what the typical FWHM is for the sub frames. If you are in a B4-5 locations I assume that this is not a pristine mountain top with sub arc second seeing so an FWHM in the order of 2.5” to 3.5” would be achievable. Check the data to see if there are any quality issues that could be due to local seeing or maybe a focuser issue.

And further on the data quality, are you using auto focus? The stars in the image of M101 are a little soft so this could be due to crappy seeing or maybe a technical issue with focusing. 

Clear skies,
Rodney

Hello everyone,

Thank you very much for your comments! Indeed, I find them all very accurate, although the topic of the relationship between image scale, resample, and SNR is still not intuitive to me and I find it difficult to grasp. I will study it in more depth. Do you say that that having a resolution of 0.7" means I need more exposure time to compensate for the lower signal capture due to my system being configured with too small pixel size? So is it true that the amount of signal I capture with tiny pixels is too low compared to the read noise and dark current of my camera, along with the noise of the sky background itself? So, would the ideal resolution be between 1 and 2" per pixel?

What does it mean that I need to "resample to somewhere between 1-1.5" per pixel resolution"? How do I do the resampling? In my optical train? Can I change the pixels of my camera, maybe by binning? It seems that with CMOS cameras this has no effect, is that right? Could it be that the only option is to reduce the focal length of my telescope? The Vixen R200SS can work at 1120 or 800 focal length. At 800 mm my resolution is 0.97" per pixel. I can't improve further. What do you advise?

Regarding the fuzziness of the stars, could it be due to the processing I've done? Considering that I used MaskedStretch, which leaves the stars in a bad shape, maybe having to touch them up caused them to appear fuzzy? I'll redo the processing (and stretch with GHS) and see if they turn out better.

Regarding color balance, I don't fully understand the concept: does it mean that the three RGB channels should be more balanced? What does this imply? Does it have to do with the histogram curves of the channels being misaligned? How do I balance them? What could be the cause of this imbalance? For color calibration, I always use SPCC, although only with broadband images, not with narrowband (NB). Of course, I use Luminance, although I have never created a SuperLuminance, which I find very interesting and will try in a reprocessing of my image to see what benefits I can obtain.

If we talk about the lack of color in the stars, what can I do to have colorful stars? Should I perhaps take very short exposures (10 seconds?) to capture stars with color and then add them to my image? Is there another option to get colorful stars?

Regarding calibration with dark frames, I will try taking some and recalibrating my exposures to see if I can achieve a slight overall improvement. For calibrating my images, I indeed use the WBPP script, which I feed with the corresponding flats and bias frames. To extract the dark current, I use the master bias. I have read that with this camera model, there is no difference between using darks or bias for the dark current. Nevertheless, I will try calibrating with darks to see if I notice any difference. It's true that I've been using my camera at -5 degrees instead of -10, but even so, the dark current is ridiculously low.

Well, that's all for now. Let me tell all of you that I am very grateful for your patience and for dedicating your time to helping people like me who want to improve our skills in astrophotography. If someday I am able to give advice to someone, rest assured that I will do it, because it is very liberating to get answers to the questions and difficulties that arise along the  fascinating path of astrophotography.

Clear skies for all!

With your RS200 your best bet is trying to achieve around 1"/px and use the extender only when the seeing is really good and with bright NB targets. This will improve your SNR by a factor of 2. You can resample any given image in PI by using Geometry>Resample. The dafult settings work very well in most cases.

Since you didn't use the RGB data for additional Luminance, and if its quality is good albeit of lower SNR, then adding it to the (using LinearFit beforehand) existing L data (averaging them together in PixelMath) will give you a nice SNR boost.

As for stars color, unless you are blowing them when capturing then you don't need to resort to that gimmick. Just separate the stars at the linear stage (after color calibration and BXT, if the case) and process them  differently to bring out colors more. I use a combination of ArcsinhStretch and HT (and possibly curves afterward) but other methods are possible too.

Darks aren't that important for the additional dark signal nowadays, which might be quite low for an IMX571 sensor, but rather for dealing with thermal noise so I'd encourage you to take darks to suit your imaging exposures lengths (with same offset and gain) at the preferred temperatures and both remove the dark signal and the noise associated with it using both ImageCalibration and CosmeticCorrection (there must be equivalent settings in WBPP but I am none the wiser).

andrea tasselli:
With your RS200 your best bet is trying to achieve around 1"/px and use the extender only when the seeing is really good and with bright NB targets. This will improve your SNR by a factor of 2. You can resample any given image in PI by using Geometry>Resample. The dafult settings work very well in most cases.

Can you help me understand this (sorry, I'm very new to all this); I do understand that if your image has a "storage" resolution of, say, .2"/px you are really not getting that level of details, because there is no way you can get that clean of a signal from where we all may live.
But my *visual* impression is that a zoomed in image with .2"/px resolution *looks* (not *is*) better than the same image resampled at 1"/px.
What is the actual visually perceived drawback of leaving the image at .2"/px resolution vs. resampling it to 1"/px?
Thanks, and sorry in advance for the likely silly question.

Yep. There is flaring on one side, clue to issue with collimation in the main. So, soft & fuzzy stars. I don't shoot mono now so hard to give an example in that regard. But otherwise examples of "proper" stars albeit RGB there are aplenty in my gallery (those taken with a standard newt).

Well, thanks a lot for all your help. I'll try to fix this issue :-(

Thank you very much,

Jordi

Dear Bob,

Thank you so much for your words! Certainly, sometimes we can lose perspective, as in the end, it's all about having fun. Indeed, I've had a lot of fun throughout the entire process, from going out with my telescope and spending the whole night out, to all the subsequent processing. The final result makes me feel proud, as it is quite cool by my standards. Even so, I believe it's always good to want to do more and better, even if it occasionally brings some frustration. With all that said, it's also important to see what your expectations are and what the reality is, because if you don't, you certainly won't have a good time. Therefore, you need to know how far you can and want to go. Hey, but nobody said it was easy, right?

Thanks a lot and clear skies!

Jordi

Hi Jeremy,

You are right, SPCC is a great tool. I use it with all my broadband images. I've never used with a NB image. I don't know why I would apply a color callibration for a false color image!

CS

Jordi

Thanks for the hint George, I'll get a try! Until now, I've used the Foraxx Script to create beautiful RGB looking stars fron NB images and I'm happy with the results.

CS

Jordi

Yes Bill, I have the Spyder X2 calibrator. The colors of the image are all the result of my personal taste. I like the blue hue of the galaxy, but I realize that maybe is too much blue there. Now I'm reprocessing with another hue… so let's see what the final result is!

Thanks and CS

Jordi