New CCD... but what calibration files / software...?

[AndyUK]

Well, after some while of debate / investigation and a lot of help from fellow members, I now have a shiny red Atik 314L+...

I know that there's some debate about calibration files (especially with DSS) and I'd been using that for some months with my DSLR trying variations on the full set (subs, flats, darks, flat darks and bias) and everything else in between, but finally came down to a set that worked for me (again, with help!). But now that I have the Atik, I'd like to try and get it right from the start (if I can) - So quite a few queries I'm afraid:

1. DARKS - As the camera has set point cooling, I'm going to set up a library at c. -10C or so. That should cover me for all but the hottest summer nights (if it should ever get over 17C at night!).

I must admit, with DSLR stacks, I'd stopped using darks and settled on flats and bias only, and as the Atik is so low on noise anyway, would it be reasonable to assume that I don't need really these below a certain exposure?

2. BIAS - Do I still take these as "shortest" exposure darks (ie 0.001s)? (Do I need them?)

3. FLAT DARKS - If using BIAS, do I need to take these?

4. SUBS & FLATS - Now I KNOW I need both of these ! The question is how to take them - I don't mean physically, but using what software...

Artemis Capture allows me to capture them, and, with looping and a B mask I can use it for focusing, and then using the sequencer, I can take create a plan for a set of exposures with lead delays, interval etc... However, it doesn't seem to have anything you can analyse them with to ensure that the exposure is more or less right (ie like a histogram on a DSLR liveview screen).

Last night, I took some tests flats, opened them up in Maxim DL (trial), converted them to .tif and then opened them in CS5 (which is home to me!) so that I could view the histogram... which I know is really convoluted, but it sort of worked last night, although I know it can't be how people work...!

So, as I see it, there's MAXIM DL vs Nebulosity 2 vs AstroArt (are there any I've missed out?). I know I'm going to have to buy one, but which one... and can anyone advise which has the best support / guides/ tutorials? DSS is VERY nice and simple, and Ideally it would be nice to have something with as simple a GUI as that, but by the sounds of it, now that I've moved into the CCD league, do I have to leave it behind(?)

(Thanking you in advance for your patience!)

[NickK]

Yup welcome to the tip of the iceberg!

I ended up using Nebulosity - it doesn't have a mass of processing features but it works well and covers both windows and mac with a single licence but it knows cameras, can stack well.

MaximDL seems to be the daddy with PixInSight offering the post processing.

[Whippy]

IMO, you won't need darks with the 314, the 285 chip doesn't need them.

I have to be honest and say I don't take Bias or flats either simply because I'm just too lazy to do it, the clone stamp tool in photoshop get a fair bit of use in my images though!

IIRC, you can use artemis capture for flats and bias. Not 100% sure on this but with the flats you expose for roughly a third of the depth before saturation. Artemis capture shows the pixel value between 0 - 65555 so it'd be around 22k I think. Best to check on that though. Lead delays?

Personally, I've only used the histogram during processing, not exposure time. Unless it's a bright object then I'll just do 10 minutes, go as deep as you can!

Why buy any other software? The only time I use Maxim is for RGB combining otherwise it's Artemis capture for data, PHD for guiding, DSS for stacking and photoshop for processing. I know pixinsight is preferable for a lot of people over photoshop and Maxim can capture, guide and stack (and more) but IMO, get used to working with the camera before giving yourself another learning curve .

Tony..

[AndyUK]

Hi Tony - To be honest, I'd hoped that darks would now be an unnecessary requirement... They're obviously not the time consuming issue they are with DSLR's, but if they're not needed, then I can forget about them (and I guess that if I ever do need any, as long as I stick to a set temp (-10C), then all I need to do is take a few and add them into the stack retrospectively...)

Ditto Bias - If I don't actually NEED them, then I won't use them (and again, if I find that I do, then it's about 5 mins to run off a few!)

I think I will take flats though (one for each filter) - Although as you say careful use of the clone tool can get rid of a multitude of sins, it's really gradients I want to get rid of (although a mixture of Gradient Xterminator and and PI DBE could also do the job as well?). However, it's nice to have the 1/3 exposure depth confirmed (with the DSLR, I try to match the histogram position to that of the subs).

(Oh - Lead delays - I meant to say lead in delays - ie a delay of 5s before the exposure starts. I currently use something similar with the DSLR as I feel that touching the remote might cause a small guiding correction to take place, so give it 5s to "calm down" again...)

Regarding software, I'd rather not buy anything at all if I didn't need to... . I'm just feeling a little lost without my histogram to tell me how much saturation I'm achieving before starting the run! To date, aside from M42, 10 mins has been pretty much my default, but due to increased chip sensitivity, I'm thinking that maybe 5 mins will be a good starting point to pick up the same signal...

I get the feeling that I may be making this WAY too complicated? (I usually do...!), but I didn't want to hamstring myself by spending £,000's on the kit and then scrimping on the software if there's significant benefit to be achieved by buying it... (That's all)

[Whippy]

I use gradient xterminator to sort out gradients but I'm coming round to the opinion that it starts to show it's limitations when you start to strech your data a little hard. I don't doubt that flat and bias frames are the 'proper' way to eliminate abberations out of your data, I just can't be bothered to do them!

Don't worry about the lead in delays, once you've got capture sorted, looping and autosaving, leave well alone!

TBH I use 10 mins subs on most things purely because I get annoyed with losing large chunks of exposure time on longer exposures if it goes wrong!

I get the feeling that I may be making this WAY too complicated? (I usually do...!), but I didn't want to hamstring myself by spending £,000's on the kit and then scrimping on the software if there's significant benefit to be achieved by buying it... (That's all)

Yes! I'm a big fan of keeping it simple. I have a relatively straightforward setup that works and I'm not about to fix what ain't broken. I don't even use EQMOD! That may sound like I'm a bit of a luddite but to me, data capture is essentially a mechanical process, the art is in the processing so if I'm happy with the data I capture, why tinker? You've already got what you need to get going so enjoy it, I think you may find yourself surprised just how sensitive the 314 is compared to your DSLR .

Tony..

[AndyUK]

Cheers again Tony - I'm sure I'll settle into a process over time - It took me a few weeks to get the DSLR imaging process sorted... and then another few weeks to get guiding sorted, so another few weeks of fumbling in the dark will be fairly familiar...!

When I've got my extension tube I'll just suck it and see... I also feel slightly luddite in not using EQMOD, but as you say, as long as you can get the subject in frame then data capture generally looks after itself (and I know I still have the joys of battling with merging channels to come...!). Sadly the weather's not so good here tonight, but I'm actually secretly glad as a) I haven't got that extension tube and I'm cream crackered from being up the last few nights... Thank God I had this week off work!

[Whippy]

No worries Andy, don't worry about doing LRGB. Once you get the hang of it, it's second nature .

Tony..

[NickK]

I'm achieving before starting the run! To date, aside from M42, 10 mins has been pretty much my default, but due to increased chip sensitivity, I'm thinking that maybe 5 mins will be a good starting point to pick up the same signal...

You'll probably run 1-2 minutes on M42! The core trapezium will end up saturating very quickly.

[AndyUK]

Yes... I took 5 mins M42 subs with the 40D but the Trap frames even at 30s was too much - I'm guessing 15s would have been better... So maybe 4s with the 285(?).

I'm looking forward to having a crack at it in November again, but probably with the Equinox 80, and maybe with a reducer - I've also been advised that it's amazing what an Ha filter can bring out so may try that as Luminence...

[roundycat]

some interesting questions there! Let's hope you can sort out the best of the answers. Have a look here Astrophotography for some advice on calibration. Ask yourself this; how many of the world's top astrophotographers take short cuts?

Your Bias frames do not need to be the shortest possible exposure. I read somewhere that one tenth of a second makes a better Bias than 1mS but I cannot for the life of me remember why. The shortest exposures you take are likely to be Flats (or the Trapezium or a PN) so 0.1s Bias frames will be perfectly ok.

Much seems to be made of Bias lately but proper calibration does not need them. There are only two reasons you might have for taking Bias, 1 is to use them to make thermally scaled Darks, recommended in AIP but not by anyone else. 2 is to use Bias masters to subtract from the individual Flats before stacking as it is quicker than shooting a lot of Darks at different exposures. This usually works ok but you might need to run a hot pixel filter as well.

Keeping a dark library is a good idea but remember the chip changes so you need to refresh the library from time to time. No shortage of cloudy nights to do that. You may find you have some partially bad columns or rows so a Bad Pixel map that you can add into the calibration is very helpful. For some reason these defects do not always come out with calibration. It goes without saying that you should dither the mount between exposures. This, coupled with a sigma reject routine when stacking, will help clean up the picture. A lot.

For my money you can do no better than buy MaxIm. It will manage the camera, guide as well as calibrate and pre-process. It does it all but stay with PS for the post processing. MaxIm will last you a lifetime so the cost/year is very small.

You will soon find out about exposure time. Check out star brightness and do not let them overload. Once you reach saturation you cannot get any star colour.

I would not use Ha for a Lum layer. It always has smaller stars and the overall effect will be to give you ugly halos everywhere. Current opinion seems to support the idea of combined RGB for Lum but you can always do away with it altogether and simply shoot more RGB.

Dennis

[Catanonia]

Andy,

welcome to the real dark side with CCD

1. Darks, on cloudy nights as suggested, take a library of them and keep them updated every 6 months or so.

2. Bias, Personally I don't bother as this is in the darks according to my reckoning if I use darks, which I always do.

3. Flats, yes of course everytime you image you should have some. dust gets everywhere especially on the MN190, i know

I used to use Maxim and DSS but have switched to PixInsight. Ok it costs £150 notes, but that and CS5 I have everything I need. The learning curve is high, but well worth it.

Plus I have found with PI, you are in complete control of your images and how you put them together, unlike DSS that is very much, bang in the files and hope for the best.

[AndyUK]

Cheers both... A lot of food for thought...!

... unlike DSS that is very much, bang in the files and hope for the best

For some reason that I can't put my finger on, this is becoming a bit of a concern for me too at the moment - It seems to have served me okay to date, but I think I would like to try out some software that gives me more control of the frame selection / stacking process...

However, I'd also like some "assistance" during capture mode - The ability to review frame saturation as the images come in. Not only do I currently use this to assess exposure time, but I also find it helps me to decide when to start and when to stop (ie when the dusk / dawn background glow is sufficiently low)

I currently have a trial version of MaximDL, so I can try that anyway (but it does seem to be HUGE!) and although I already have Pixinsight LE, I only currently use it for DBE purposes, so I don't really know what else it can actually do . It looks as though I'll have to do some research pretty sharpish to get the most out of the Maxim trial and also try and get my head around the steep learning curve of trying to learn both it and PI so that I can make a good comparison. However, having already bought CS5, I certainly don't want that to be a "wasted" investment!

Oh and Dennis - Thanks for advice on not using Ha as luminance... Would some people sometimes perhaps add a %age of Ha into the luminance...?

I must admit I'm getting a little confused about the use of the L filter at the moment, as surely it would also capture LP(?). The idea of making the L layer from RGB (with the Na/Hg spectrum gap between G and sounds very logical to me... The only other solution I thought of surrounding this issue was to swap out the L filter for an LP filter(?), but I guess they also have a colour cast to them which might make colour balancing tricky later on(?)

[NickK]

If you're using LP for all filters in the filter wheel then you can put the LP filter on the nose of the adaptor. The light then passes through the LP filter as it enters the nose of the adaptor then hits the filter wheel filter (LRGBC) before hitting the camera sensor.

[adamsp123]

Dunno if this is any help, being new to cooled CCD mono (Atik 383L+) I have made a darks library, with set point cooling it is easy, BTW I have plumped on -20C, I take flats, I need them, then I stack each of the LRGB individually in DSS using the lights, flats and darks with DSS set to mono mode, it works for me as I am familiar with DSS and it is easy to use.

I align each of the binned colour to the Luminance in Registar, again a very simple program to use, then process in PS..... it is early days for me but at least by using this method I can get something without spending a lot of time learning some new process.

[roundycat]

Andy, pm on its way

Dennis

[davew]

Interesting thread.

I've downloaded much of the Software mentioned and find them all good in their own way.

DSS has a major flaw. It's so bad that it will never gain the acceptance it deserves ( Which is it works ). The fix is easy but it won't be changed any time soon I don't suppose. That flaw ? It's free. Can't be any good can it ?

I have yet to hear a properly reasoned argument as to why it won't work well for the vast majority of people in the even vaster ( ! ) amount of cases. What deal breaker is missing ? It even has it's own version of a bad pixel map. Trouble is the author calls it " Cosmetic ". If you think it blows holes in all your over exposed stars, have a go with any of the others, put in equal parameters and see the results.

That said, I've tried CCDStack and find it really nice in it's graphical representations. You get to see a good picture of cut pixels. It also has the only decon method I've had any success with. Problem is the price.

Dave.

[AndyUK]

Cheers all...

Nick - I understand what you're saying, but I do like the logic of one of Dennis' suggestions of creating the L from more RGB's...The gap between R&G will effectively give me the LP reduction result I need at no extra cost other than time having to take more RGB's. I think I'll try the "low cost" route first and if that doesn't work, then I can look at trying the way you've suggested - Again, it makes sense, although the con is that it would be adding another optical element into the light path(?).

Adam - As DSS is free (and I have it, and I know how to use it), I'll quite possibly start off trying to use that first - As I think I mentioned earlier, I know I have a habit of trying to over-complicate things and certainly I'm always impatient to spring well before I can crawl... I need to save up for filter wheel and filters as well, so software might have to wait a while longer yet (although Registar is also very high on that list ). Thanks also for the tips on slight changes in DSS - When I can finally take some CCD subs, I'll give it a whirl (although I only have UV/IR cut and Ha filters that I can use at the moment...)

Dave - I hear what you say - I remember something similar in days at Uni... If you had to pay 50p to get in to the residence bar to see a band, it was worth going and the place was packed, but if it was free, even with the same band(!), the place was empty. I guess not always does the saying "you get what you pay for" hold true (although I have to save from personal experience most of the time it does).

It's beginning to look as though choosing software is going to be more complicated than choosing the camera...

[davew]

I know what you mean Andy. If he charged £ 20 for it, a lot less would use it and less would damn it.

From what I've tried and used I like Maxim for capture and guiding ( sometimes ! )

PHD for guiding ( other times )

CCDStack for stacking Decon and DDP.

Pixinsight for imaging specifics.

Photoshop for all else. Especially layers.

It's a shame I can't afford them all.

Dave.

[ollypenrice]

You can't do M42 with one exposure time. Some use two, some three. I used three. I thnk they were 15, 50 and 300 secs an the 4000 chip at f5.3. Then you need this excellent tutorial to put them all together.

Catching the Light: Astrophotography by Jerry Lodriguss

That gave me this result:

http://ollypenrice.smugmug.com/Other/Nebulae-and-clusters/M42CCBOV2010/1100345185_HHd4m-X2.jpg

A technique I use a lot is to do two or more stretches of the same data, eg one for galaxy cores and one for outlying nebulosity and combine them as above. Very careful control of levels and curves might get round this but I go for the multiple stretches approach.

I'm with Dennis on another point, too; the history of my imaging is easy to tell. Stop taking any shortcuts, ever!

Olly

[AndyUK]

Dave...

... It's a shame I can't afford them all

Hmmm... Yup... There's the rub... probably for everyone! If only I had more money... and even then we'd then probably need just that little bit more . I think I may well have asked a question the answer for which I can't afford...

(Olly... I hear you... Loud as a bell... !)

[roundycat]

Andy, let me have your e-mail and I can send a full set of answers to the points you raised earlier. Apparently it is too long for the forum mail service.

Dennis

[NickK]

Cheers all...

Nick - I understand what you're saying, but I do like the logic of one of Dennis' suggestions of creating the L from more RGB's...The gap between R&G will effectively give me the LP reduction result I need at no extra cost other than time having to take more RGB's. I think I'll try the "low cost" route first and if that doesn't work, then I can look at trying the way you've suggested - Again, it makes sense, although the con is that it would be adding another optical element into the light path(?).

Correct.