A Quartet of interacting galaxies about 70-75 million light years distant with a recession velocity of ~ 1,600 km/sec.
In spite of this a few photons managed to penetrate through my light polluted back yard sky and here is the result. It proved to be one of the hardest challenges I've tried so far. I searched for the "best" sub lengths but this effort was defeated by constantly changing conditions from night to night. In the end I just took all the varying subs, different lengths and all, and let CCDStack combine them.
Luminance = 280 min total at .88 arc-sec through a Astronomik CLS and the TEC180/G2-8300 CCD.
Colour proved more difficult as I'm having issues with blue fringing with the Trius/TEC140/Baader filters. Some of the blue was through the Hutech IDAS and some was through a Baader Semi-Apo. The G & R were unfiltered. I consequently had some bad gradients and in retrospect I should have used the Hutech filter for all colour. And, I wish now I had exposed a lot more blue. With the Baader filter, which gives quite tight blue stars, I probably need 30-40% more time compared to unfiltered. Colour subs ranged from 300, 420, and 480 seconds: R: 116 min G: 116 min B: 130 min
Thanks for looking! Your helpful comments are always instructive and appreciated!
hi Peter. looking pretty impressive overall, with only a bit of noise appearing to be the main remaining problem.
As I understand it, apart from spectral filtering, the two approaches that work in pollution are:
1. optimise the sub length for the conditions (if it varies, err on the longer side)
2. use some very high quality flats (also need high quality darks (and possibly bias) to calibrate them). What sort of flat procedure do you use?
Once you have done that, there is nothing but extra total exposure time - each extra mag in sky brightness requires very roughly double the total exposure time - but more total time will certainly do the job if your flats are good enough. Longer subs will not help though.
That's actually a nice image Peter. I don't see any troublesome blue halos here so it looks like you are pretty much on top of that.
Yes longer total exposures is the way to go. Even with my 17 inch 8 hours total is a bit on the light side. 12 hours is a good target. Judging from many of the images here and Rob Gendlers original works 12-30 hours is good depending on the aperture and dark skies, camera sensitivity etc.
I thought the strategy with CCDs was shorter exposures with light pollution ( the sky glow builds up faster than the signal) and longer with dark skies.. Also with the Trius short exposures are a good solution to the haloing effect anyway as Ray's images show.
That's actually a nice image Peter. I don't see any troublesome blue halos here so it looks like you are pretty much on top of that.
Yes longer total exposures is the way to go. Even with my 17 inch 8 hours total is a bit on the light side. 12 hours is a good target. Judging from many of the images here and Rob Gendlers original works 12-30 hours is good depending on the aperture and dark skies, camera sensitivity etc.
I thought the strategy with CCDs was shorter exposures with light pollution ( the sky glow builds up faster than the signal) and longer with dark skies.. Also with the Trius short exposures are a good solution to the haloing effect anyway as Ray's images show.
Greg.
Thanks Greg!
I guess I will need to try imaging over more nights. With the weather so erratic and my inability to see very much to the East I'm limited to picking targets that are overhead pretty much. As the evening progresses I then get gradually worse and worse LP as I track to the West. If I get 3 r 4 nights in a week or two I'm inclined to start processing! I will need to become more patient! I'm around 10 hours total if I include RGB time so perhaps not too far off, but as Ray points out the number of subs required starts to increase dramatically if I want to do better.
By the way, how do you deal with a large number of subs, say luminance with something like 40 subs? Would you stack them all in one go? And, do you ever combine differing sub exposures? I did in this case and just let CCDStack put in a weighting factor.
hi Peter. looking pretty impressive overall, with only a bit of noise appearing to be the main remaining problem.
As I understand it, apart from spectral filtering, the two approaches that work in pollution are:
1. optimise the sub length for the conditions (if it varies, err on the longer side)
2. use some very high quality flats (also need high quality darks (and possibly bias) to calibrate them). What sort of flat procedure do you use?
Once you have done that, there is nothing but extra total exposure time - each extra mag in sky brightness requires very roughly double the total exposure time - but more total time will certainly do the job if your flats are good enough. Longer subs will not help though.
Cheers, Ray
Hi Ray,
I appreciate your comments as always.
Firstly, I have not figured out how to dither two scopes at the same time especially since I'm often taking different length exposures. But, even with the same exposure I can't quite see how to do it given the differing download times of the two cameras (the Trius is fast and the G2 slow). But, I know dithering would really help the noise so I'd like to attempt some sort of solution.
As for darks, bias. flats. Currently I'm taking 40 bias, 20 darks, and 50 flats. The flats have a mean adu of 33,900 with a min= 9,600 and a max = 52,158. Should I be looking for an mean that is lower given that max value? I've generated all of my calibration frames in CCDStack. For rejection during stacking I use the Standard Sigma Reject for luminance when I have a lot of subs, and the Poisson reject for RGB. If you have another suggestion please let me know.
So, if I've got this right the issue in LP is the decreasing (with time) of the dynamic range of the camera. And, it's the dynamic range that allows the real signal to be seen over the noise. But, I think I can safely say that 10 1 min subs will not give the depth of a single 10 min sub. Getting the balance just right between those two extremes is the tricky bit.
Looks beautiful to Me, Peter.
Splendid work.
Cheers
Ron, you are kind to say that! It is a beautiful subject to work with for sure!
Peter
Quote:
Originally Posted by astronobob
Looks Grouse Peter, them galaxies are really popping in this view
Bob, I had to look up the word "grouse" and was quite pleased to learn that you had not misspelled the word "gross!" I guess that reveals my non-Australian origin. Hey, I'll take it. Thanks very much!!
By the way, how do you deal with a large number of subs, say luminance with something like 40 subs? Would you stack them all in one go? And, do you ever combine differing sub exposures? I did in this case and just let CCDStack put in a weighting factor.[/QUOTE]
In the case of my Proline 16803 the files are 32mb each so I tend to process in batches of 10 subs. My computer would not handle 40 subs.
I then stack the resulting combines of 10 files each to make a final master.
As far as differing lengths, firstly, I tend not to do that, but I probably have a few times. I made no allowance at all and simply added them into the combine. I suppose a weighting makes sense. For me that would usually be 10 and 15 minute subs. Now that I think about it I did them separately, all the 15minute subs into a master, all the 10 minute subs into a master. For one thing they would have used different dark masters for dark subtraction. I find though with the FLI cameras and the SX Trius dark subtraction is not that fussy except for when I am using the CDK17 in which case it seems to make a huge difference to the success of the flats for some reason I don't fully understand.
Generally with a large number of subs it makes sense to flick through them and watch star sizes, guiding, perhaps gradients, clouds, satellites, wind affected and weed out the lousy ones. Then combine after callibration,registration, normalising and data rejection.
A few lousy subs can worsen. 30 subs combined may make a nicer image than 40 combined with 10 poor subs.
Peter, it is my favourite cluster of galaxies in the Southern Sky, I observe it on numerous occasions during it's favourable viewing times.
I have been checking this group for Supernova for over 15 years.
Again,lovely presentation.
Cheers
Hi Peter - looking good.
Did you stretch the 32 bit stacks in FITS Liberator with the x^(1/5) function?
( maximum compression )
It might make the transition from dark background to galaxy edges much smoother on the black point.
Maybe you would see an improvement ?
Hi Peter - looking good.
Did you stretch the 32 bit stacks in FITS Liberator with the x^(1/5) function?
( maximum compression )
It might make the transition from dark background to galaxy edges much smoother on the black point.
Maybe you would see an improvement ?
cheers
Allan.
Hello Allan,
No, in fact this is the first I'm hearing about this technique. Might you explain a little more about how this works and how/if it fits in with a processing sequence using CCDStack? Would I just import the fits out of CCDStack unstretched after stacking?
No, in fact this is the first I'm hearing about this technique. Might you explain a little more about how this works and how/if it fits in with a processing sequence using CCDStack? Would I just import the fits out of CCDStack unstretched after stacking?
Thanks!
Peter
Yes - & place the the 32 bit files into NASA FITS Liberator for stretching.
Firstly, I have not figured out how to dither two scopes at the same time especially since I'm often taking different length exposures. But, even with the same exposure I can't quite see how to do it given the differing download times of the two cameras (the Trius is fast and the G2 slow). But, I know dithering would really help the noise so I'd like to attempt some sort of solution.
As for darks, bias. flats. Currently I'm taking 40 bias, 20 darks, and 50 flats. The flats have a mean adu of 33,900 with a min= 9,600 and a max = 52,158. Should I be looking for an mean that is lower given that max value? I've generated all of my calibration frames in CCDStack. For rejection during stacking I use the Standard Sigma Reject for luminance when I have a lot of subs, and the Poisson reject for RGB. If you have another suggestion please let me know.
So, if I've got this right the issue in LP is the decreasing (with time) of the dynamic range of the camera. And, it's the dynamic range that allows the real signal to be seen over the noise. But, I think I can safely say that 10 1 min subs will not give the depth of a single 10 min sub. Getting the balance just right between those two extremes is the tricky bit.
FWIW, I used to use 25 darks with my H694, but found that it needed a lot more - dark noise can be imprinted on the flats and that can lead to noise that does not reduce with extra exposure. With heavy LP, I would guess that at least 50 darks would be needed - maybe even more. I would also use maybe 100 bias and flats (33900 should be fine for the flats, but 52000 may just possibly be getting a bit close to the non-linear region of the CCD). If the LP is severe, you might even get some advantage from even more flats. Of course dither will fix most of the noise if you can get it going.
FWIW, I used to use 25 darks with my H694, but found that it needed a lot more - dark noise can be imprinted on the flats and that can lead to noise that does not reduce with extra exposure. With heavy LP, I would guess that at least 50 darks would be needed - maybe even more. I would also use maybe 100 bias and flats (33900 should be fine for the flats, but 52000 may just possibly be getting a bit close to the non-linear region of the CCD). If the LP is severe, you might even get some advantage from even more flats. Of course dither will fix most of the noise if you can get it going.
Hi Ray,
Thanks for the MaxIm tip. I'll have a look.
OK, I see I will be spending a lot of time doing darks. I have not bothered doing any darks/flats for the Trius, only for the KAF8300 chip.
I'm going to need to redo my flats anyway as the camera is coming off for a good cleaning!
What an awesome field of galaxies. I dig the composition; weighted to either side of the frame and the processing is lovely.
H
Thanks H. I can't take much credit for the composition as it was the only way I could fit all 4 in the same FOV! I'm glad you think the processing was "lovely." I certainly didn't feel that way doing it!!!
Hi Peter - looking good.
Did you stretch the 32 bit stacks in FITS Liberator with the x^(1/5) function?
( maximum compression )
It might make the transition from dark background to galaxy edges much smoother on the black point.
Maybe you would see an improvement ?
cheers
Allan.
Allan, I gave this a bit of a tryout and yes the transitions do look quite smooth in comparison. I will need to get more experience with this program. Thanks again for pointing it out to me.
Allan, I gave this a bit of a tryout and yes the transitions do look quite smooth in comparison. I will need to get more experience with this program. Thanks again for pointing it out to me.
Peter
Thanks Peter,
I always stretch the 32 bit FITS file with NASA FITS Liberator
& then save it as a 16 bit Tiff file for Photoshop.
I have found the x^(1/5) function is the best -
it offers the most compression.
Be careful of the left & right sliders.
You need to move the left slider to just get rid of the blank area
or fully left -
& the right slider so you don't clip the brightest stars.
( which will appear as vertical lines)
It will then need further stretching in Photoshop but not much.
It's a powerful function that does most of the stretching work for you using mathematics.
You can even experiment with different functions depending on the picture.
ArcSinh(x) can be good for nebulas.
x^(1/5) can really help in those tough areas where you're trying to pull
galaxy arms out of the noise in the black point.
