Calibrating 2x2 Binned Images

[gnomus]

I've finally managed to collect some data with my new CCD camera. I shot a series of Luminance shots at 1x1. The RGB data was shot at 2x2. I took some flats, bias and darks all at 1x1. I've been able to calibrate my Luminance data. I am now unsure about how I calibrate the 2x2 data. Do I need to shoot new bias and dark frames at 2x2? (I can do that.). Flats I am stuck with since I've stripped the rig down.

Can someone give me a couple of pointers?

[ollypenrice]

Certainly both darks and bias need to be binned as per the lights. They are recording camera noise in whatever mode you're shooting. You can easily shoot some 2X2 darks and bias so that's not a problem because no light is getting to the chip. Do them any time. I suppose that, ideally, flats should really be shot as per the lights they are to calibrate but, since a flat is just a photo of the imperfections of your lightpath, it should be possible to resize a 1X1 flat downwards to make it fit a 2X2 light. I'm pretty sure this will work. (Professional flats contain pixel by pixel information but ours don't.)

If using set point cooling you only need bias to use as darks for flats. If you shot your flats 1X1 then you don't need 2X2 bias.

Olly

[gnomus]

OK - I built my darks and bias frames at 2x2 binning.  I couldn't get flats to work.  I did however get an image of the Pelican that - superficially - looked OK.   Problem is when I looked close up, a lot of my stars had strange artifacts.  Like this:

Any idea what could be giving rise to this?  This is just a stretch and a crop - no noise reduction or sharpening...

[ollypenrice]

Never seen that before but it could be a lack of anti-aliasing in your stacking settings. Anti-aliasing tries to break down edge artefacts created by high contrast. On the other hand I've no idea where or why this might be happening so it isn't a very useful post!

Olly

[gnomus]

Never seen that before but it could be a lack of anti-aliasing in your stacking settings. Anti-aliasing tries to break down edge artefacts created by high contrast. On the other hand I've no idea where or why this might be happening so it isn't a very useful post!

Olly

Jings, Crivens, Help ma Boab.  If you don't know, Olly, I guess I'm really scr*w*d!  

No.. seriously, I am grateful for your thoughts...

I have been calibrating and stacking in PixInsight - following a tutorial that was recently posted on this site - http://lightvortexastronomy.blogspot.co.uk/2015/07/tutorial-pixinsight-pre-processing.html

I'm not sure I've seen an anti-aliasing option in the few (DSLR) images I have stacked in PixInsight - this is my first go at CCD imaging - is antialiasing called something else in PI? 

I should also point out that when capturing I followed a "recipe" I found in a book.  That chap used different equipment and I think I ended up underexposing quite markedly.  For my Pelican shot I set up 20 x 90 seconds of Red (used as luminance), followed by only 8 x 90 secs of R, G & then B (binned at 2x2).  At the end of the night I had a few minutes left over and threw in another 5 x 180 seconds of 1x1 red to use as luminance (I think I realised I'd underexposed by that time).  15 Flats later I had to strip the gear down and go get some beauty sleep.  In the morning, I did find that I had to stretch my images quite considerably to get something on screen.  

[ollypenrice]

Why are you using red as luminance? Luminance isn't red, it's the full visual spectrum. Planetary imagers use it as luminance because it beats the seeing better than shorter wavelengths but this strikes me as an abomination in deep sky imaging. At one time many people (including me) used to use Ha (a special case of red) as luminance and produced pink images with blue star haloes. Happily we've learned a lot since then! There is a temptation to use some Ha as luminance on emission nebulae which are Ha dominated but even then it should be done judiciously.

I'd use luminance as luminance for the same reason that I'd use a hammer as a hammer - because that's what it is! 

Olly

[gnomus]

Why are you using red as luminance? Luminance isn't red, it's the full visual spectrum. Planetary imagers use it as luminance because it beats the seeing better than shorter wavelengths but this strikes me as an abomination in deep sky imaging. At one time many people (including me) used to use Ha (a special case of red) as luminance and produced pink images with blue star haloes. Happily we've learned a lot since then! There is a temptation to use some Ha as luminance on emission nebulae which are Ha dominated but even then it should be done judiciously.

I'd use luminance as luminance for the same reason that I'd use a hammer as a hammer - because that's what it is! 

Olly

Thanks again Olly. I have been responsible for many abominations in my time, but this may be my first truly astronomical abomination.

As I said, I was following a 'recipe' in a book I read ('The 100 Best Astrophotography Targets'). The recipe suggested that I take 83 x 1 min Red as luminance, then 20 x 1 min binned R & G, and finally 27 x 1 min binned B. (I wasn't able to get that amount of data and so scaled everything down a bit in an attempt to get something in one evening.). As I am just starting out I am trying to get advice as to what the easiest targets might be at this time of year.

I think the author's idea of multiple short exposures was an attempt to stop the brighter stars from blowing out.

I can try the Pelican again, using luminance as luminance and trying longer exposures. If, however, anyone has a better suggestion for a beginner target at this time of year, I would be happy to switch.

[Davey-T]

I've just been using a screwdriver as a chisel and hitting it with a big spanner, is this why my astro' images are [removed word] rubbish ?

Dave

[gnomus]

I've just been using a screwdriver as a chisel and hitting it with a big spanner, is this why my astro' images are [removed word] rubbish ?

Dave

From what Olly is telling me, it sounds like I'm the only tool here.

[ollypenrice]

You do come across some rum ideas in this game and many will think mine are as rum as it gets. Fair enough. However, I think that taking '83 x any number' of DS subs is as close to howling at the moon as we are likely to see on here. The law of diminishing returns sets in somewhere between 30 and 40 subs. Eighty three? Why eighty three? Damn it man, eighty seven would be so much better! Well, no.

If you have a shallow well depth camera then short subs for star colour might be an idea but I feel happy enough with my star colour in 15 minute subs, in this case at F 6.8.

https://ollypenrice.smugmug.com/Other/Best-of-Les-Granges/i-VTWp94c/0/X3/M27_HaO111RGBWEB-X3.jpg

The need for short subs is bandied about endlessly but, in my view, the real need is to control the stretching of the linear data properly.

Olly

[glowingturnip]

I've just been using a screwdriver as a chisel and hitting it with a big spanner, is this why my astro' images are [removed word] rubbish ?

Dave

yeah, that's how I do my collimation too

I've got the '100 best AP targets' book too and use it for ideas for new targets (though I do try not to limit myself just to those).  I have to say though, ahem, that his pictures aren't as good as many on this board, so I tend to take his capture and processing tips with a pinch of the proverbial

[gnomus]

OK - I built my darks and bias frames at 2x2 binning.  I couldn't get flats to work.  I did however get an image of the Pelican that - superficially - looked OK.   Problem is when I looked close up, a lot of my stars had strange artifacts.  Like this:

starsexample.jpg

Any idea what could be giving rise to this?  This is just a stretch and a crop - no noise reduction or sharpening...

Apologies for this post.  I was rushing ahead with things in PixInsight and I think that I combined my L, R, G & B channels incorrectly (ie all at once).  I reprocessed the image doing a Channel combination and then adding luminance (as per Harry's videos), and I didn't get those strange artefacts in the stars.  This is my first image with the CCD.  Clearly, I didn't gather enough data, but it is significantly less noisy than my DSLR images, which is encouraging.  The colour balance leaves something to be desired.  I wonder how one decides upon the relative amount of red, green and blue data that one should gather for each object.  Anyhow, I'm not going to spend too much time working on the image - I was pleased to get anything at all - but I'll gather more appropriate data next time and devote what energy I have to working on that.  Thanks again for the help.    

Steve

[gnomus]

I had another go last night (22 July) using the correct filter for luminance.  I also went for 5 minute exposures (but think I could have gone for even longer).  I then combined last night's data with the stuff I got a few nights ago.  

For anyone interested in how this mess turned out, I've posted the result here: http://stargazerslounge.com/topic/248909-first-attempt-with-new-ccd-pelican-nebula/