Out of focus image suggestions

[MattJenko]

Hi,

Bit of a daft question really, but I am just starting out using a mono camera and filters and from my first few sessions I have a few colour sub sets that are out of focus, and the old adage of beggars can't be choosers applies. I have been imaging open clusters to make sure I can get stars right so there is no smeared nebula or galaxy to give up on, just some out of focus stars. I have taken this as an opportunity to try and practice processing them, as not all the channels are out of focus, and I have been taking the in- focus channel as a luminance channel for the overall RGB image, which is kind of working.

What suggestions does anyone have on 'refocusing' individual colour channels for pure star fields, basically making the stars smaller I guess? I use PixInsight and have been trying the various MorphologicalTransformation options with some limited success and will give deconvolution a go a bit later today, just wondered if anyone had any other fancy suggestions? I am not aiming for APOD, just some different techniques to practice on my poor quality data.

Thanks

Matt

[ollypenrice]

I tend not to use the luminance layer for the starfield where possible, nor do I shoot lum for clusters - though this means your colour has to be unbinned and in focus. (In focus lum will greatly help out of focus RGB.)

I make my starfields, then, from RGB only and stretch using a specific curve which tries to get the background sky to the same level as the main image (23/23/23 as measured in Ps for each channel) while stretching the stars as little as possible to keep them small and colourful. This is a curve designed to control Alnitiak but the idea is similar. The slight droop in the straight line is a boob! Sorry. Should be straight.

To blend the 'stars and background sky' image with the main one I paste it on top of the main image and fine tune the background sky till it is identical to the main image background. This is vital. Then I set the opacity of the top (stars and background) layer to zero to reveal all the faint fuzzies beneath and I run the feathered eraser over them with the top layer active but invisible. When the top layer opacity is restored you have the RGB only stars and background with the faint fuzzies coming through from below.

Although this M31 had about 50 hours the outlying starfield (beyond the massively stretched outer halo) is made from a trivial quantity of RGB only. Less than an hour per channel from memory.

Olly

[MattJenko]

Thanks Olly. The only salvation (other than to retake the G+B data which I will at some point, but forecasts are gloomy) has been the using the R channel as a Lum channel and combining after the RGB nursing and combination. The results on the 2 open clusters I took that fateful evening are not bad considering it is otherwise throwaway data. I used PixInsight, but did something very similar to what you suggested within that application.

The images are still far from ideal and it is a great lesson for me about the importance of focus and concentration, and the acceptance that good images from mono CCDs are going to be nearly always multi evening affairs.  (M36 and M37 by the way).

That M31's stars are wonderful by the way. As well as not using much exposure time, and doing RGB only, do you rely on the quality optics of your rigs to get them so crisp - ie how do you get them so sharp?

[ollypenrice]

That M31's stars are wonderful by the way. As well as not using much exposure time, and doing RGB only, do you rely on the quality optics of your rigs to get them so crisp - ie how do you get them so sharp?

Obviously the optics don't do any harm but their 'expensive aspect' is, above all, their ability to cover the full frame Kodak chip without distortion. Less expensive optics will often match the Taks but not at the same F ratio and not over such a large image circle - but if you have a smaller chip the latter may not matter at all. As for the former, you just take a bit more data. I honestly think that the critical things are focus and the nature of the Curve you use to stretch the starfield.

I don't know how much Tom paid for his half of the tandem but my FSQ106N, quite an old lady, cost less than £2000 in the end. Not free but not Ferrari either!

As for multiple nights, I've argued many times that a mono CCD is the fastest imaging system available. It just is! The trouble is that when folks start using them they set themselves higher standards and find they want to spend more time.

Olly

[MattJenko]

Thanks. Focus has been successfully drummed into me after this experience!

I have read your points about mono being quicker, and I agree in principle, but I think an additional factor for mono beginners like me is that there are many more actions to take during the evenings data gathering: filter changes, refocus, reframe (if focus was done on a star elsewhere), and options on how to take an image (all one channel or R,G,B,R,G,B...) that it will take me a while to get up to that speed and routine of being able to make mono imaging a faster route than just setting up the DSLR or OSC for the evening. Processing is another area where it will take me a while to streamline to be faster than my DSLR processing.

You are absolutely right in that my own standards have changed very quickly, the arrival of an Ha filter last week is testament to that

[ollypenrice]

Scientifically it is best to shoot blue and lum as high as possible and use the low elevations of the target to catch red and green. Psychologically (ie to reduce the potential for suicidal despair...  ) there is something to be said for setting an electric wheel just to scroll LRGB, LRGB etc.

Olly