Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Stargazing Processor power for entry level Astrophotography

  1. Sep 13, 2016 #1

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    I saw what a fellow member of my Astro Society had obtained from fairly basic gear and I reckon I should be able to 'afford' something like he has. He is using his elderly DSLR, mounted on an ioptron startracker and is producing pictures of nebulae which I couldn't hope to obtain by eye from my albeit semi rural location near London. Using a stack of images and pasting them together is clearly a good method for noise reduction but would have to do that indoors on my iMac.

    Two problems here - there are relatively few suitable OS X programs available (for cheap or for free) so I would consider buying an extra machine - a Windows laptop. The very cheapest Windows laptops are really cheap but I have a feeling I could be all night processing a dozen pictures with such a low performer. In terms of i3, i5 and i7 grades, where should I aim?

    Is Nebulosity a good prospect? I must say, I do not find it at all intuitive and the Licence is around GBP 90, which is not a trivial amount if I don't get on with it any better than with the demo version. Is it worth sticking with or is there something free that's as good on a PC?
     
  2. jcsd
  3. Sep 13, 2016 #2

    Grinkle

    User Avatar
    Gold Member

    For image processing, if you are not swapping, you will see your performance scale with processor. If you are swapping, you will be having a sub-optimal user experience, is a polite way of putting things. If you don't have enough RAM to avoid swapping, spend on RAM before processor. Personally, I would say an SSD over a mechanical drive is potentially a higher budget priority than processor generation, but maybe that is not so true for image processing if you already have plenty of RAM. Where to aim is a question of budget.

    I like this site for articles on bang-for-the-buck vs different build options.

    http://www.tomshardware.com/t/hardware/

    Regarding software, I haven't heard of what you mention, can't comment.

    You might consider downloading a 30 day fully functional free trial of Adobe stuff and if you like it, subscribe for a few months to experiment with your images, especially if this is a thing you will potentially only be doing for a few months. Adobe has very highly functional image processing software with lots of internet based expertise available for help / learning. (I use Adobe, I am not otherwise affiliated with them).
     
  4. Sep 13, 2016 #3

    Chronos

    User Avatar
    Science Advisor
    Gold Member

    Processing power should not be too much of an issue. While APT and Nebulosity are quite good and quite expensive, imo, I use APT [astrophotography tool] which works well for DSLR users and is inexpensive.
     
  5. Sep 13, 2016 #4

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    That's very true and I've been there with every computer I have owned. Buying and fitting extra RAM is easy and cheap and it's something I've done often. More RAM can be like getting a new computer.

    I'd like to start off with the 'right' decision. I guess what I'm after is a flavour from PF members who've actually experienced limits. I do a fair bit of Photoshoppery and a slow machine, even without swapping, can be a drag. I know OS X machines can be noticeably slower than Windows so, if I end up with a Windows laptop, I may well be pleasantly surprised - as long as the required grunt is not too great. As I understand, it will involve multiple (tens of) images (perhaps 20MB each), and some long winded operations. But it isn't real-time processing so you may be absolutely right. I don't want to spend £1k when I can get away with £300.
     
  6. Sep 13, 2016 #5

    davenn

    User Avatar
    Science Advisor
    Gold Member

    laptop is good if you want portability
    if you are pretty much only going to do the processing indoors at home, then a reasonable AMD i5 or i7 desktop (or mini tower)
    will keep you out of trouble and will be 1/2 the price as a similarly powered ( cpu/ram) laptop
    memory is much more important ... you want a minimum of 4 gig and preferably 8 gig ...... image processing is memory and
    cpu intensive specially when doing stacking of many large file ( RAW ) images

    Nebulosity 3 is a VERY good prog, I have had a play with the demo version 2 down sides are very expensive for me AU$250+
    and it has a very steep learning curve ... no probs for a sharp mind and someone with lots of time to learn

    Deep Sky Stacker is free and if you already have Adobe Lightroom or Photoshop, then you are in business
    the Adobe bundle is very affordable these days on subscription ... for me ~ AU$14 / month and all the updates come with that
    never again is there a need to purchase a full updated version at many $100's as you used to have to do in the past


    Dave
     
  7. Sep 14, 2016 #6

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    I have a 3GHz i3 iMac with 8GB of memory 'indoors' and I am unlikely to be improving on that. The screen is truly wonderful, in the position I use it and I doubt that I could improve on it with a non-Apple monitor without spending Apple prices.
    I expected that reaction. :frown: I will just have to bite the bullet, I guess.
    Encouraging news is that I took six images (8s each) of the Pleiades (still had oval stars of course) and managed to align and stack them in Nebulosity 4 demo. The noise was visibly (just) improved, compared with just one of the images and some stars were revealed after the processing. So that has made me realise that it will be worth while. The price of N4 seems to be about 90USD. DSS is only for Windows, unfortunately.
    I recently bought a new PS Elements (not rented) and that hurt a little bit but I have not yet come to terms with a monthly fee for using the rest of Adobe's stuff.They have you by the throat, once you are committed to using a package. OK if it is paying for itself for a pro user but I may go for a month without working on pictures.
    PS Presumably the noise reduction process in a camera is not optimal for astro imaging but how where would the ideal exposure time / number of images lie? I can only do 30s with my k10D but that will require a star tracker, of course.
     
  8. Sep 14, 2016 #7
    I have no experience of astrophotography, but I wonder if you could not get the results you are after simply using Python.

    Modules such as SciPy, numpy, PIL and OpenCV provide excellent functionality for image processing. It would have the added benefit of being free, and suitable for your iMac. Just a thought.
     
  9. Sep 14, 2016 #8

    davenn

    User Avatar
    Science Advisor
    Gold Member

    I only mentioned DSS because of your comments of possibly getting a Win based machine
    I haven't had a look at Neb4 to compare with Neb3 for differences in price and features

    from the www site, that's around AU$130, not too wicked

    I originally started with PS Elements 11and a stand alone version of Lightroom a number or years back but its limitations over the full version of PS
    and the aging version of LR meant that upgrading to the new subscription that includes both of them was a "no brainer" As I would going to have to
    spend $100's to individually upgrade both of them.
    I use LR the most, it gets used for ALL my raw file editing and does an outstanding job. PS gets used for the special actions that LR doesn't do

    Speaking of RAW files, if you are not shooting in RAW, START doing so immediately for all your photography.
    The last thing you want to be doing is trying to edit or stack and edit jpg images.


    most of us do very minimal NR in camera, keeping the setting at the lowest level. The huge disadvantage of in-camera NR is it extends the time to save
    an image after taking it and that becomes really frustrating real quick.

    Basically ... the longer the exposures and the more images continue to improve the stacked result. Not sure when the point is reached that more
    images = diminishing returns ?. Many of the top astro-tog's have no qualms of stacking 100 x 2 minute exposures and the results are amazing.
    I am often doing between 10 and 20 x 30 sec images stacked and I can see that extending my exposure time is a definite must.
    I am very keen to get out to 2 - 5 minute exposures and stacking 20 - 30 of them.

    eg. There is just no substitute for longer exposures, 60 x 1 sec exposures stacked will NEVER equal the quality of a full single 1 minute exposure.

    most dSLR's can only do 30 sec max exposures in-camera. For longer than that you buy a shutter release cable/controller
    not the one I have ... I have the same make, just a different model. Essential for doing exposures longer than 30 sec
    http://www.digitalcamerawarehouse.com.au/prod12892.htm


    cheers
    Dave
     
  10. Sep 14, 2016 #9

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    I changed to RAW as soon as I bought the k10D. As far as I'm concerned, JPEGs are a bit like MPEGS. They are OK when used in context but they aren't as good as the real thing. :smile: They do eat up space on the card and on the HD, though. But I am good a chucking most of my failures, rather than keeping them on the off chance that a miracle can turn them into a masterpiece.
     
  11. Sep 14, 2016 #10

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Thanks for the suggestion but I think that stuff's a bit too low level for my use. I stopped using command line a long time ago and I suspect there may be quite a lot of that in your suggested programmes (?). Mr Jobs first tempted me into user-based stuff (when MS Dos and Unix were all that was available) and I do like to have my hand held when I want to achieve a result. No excuse, I know and I am probably missing out.
     
  12. Sep 14, 2016 #11

    Andy Resnick

    User Avatar
    Science Advisor
    Education Advisor
    2016 Award

    IMO, you should be able to use whatever gear you have on hand, there's no substitute for your brain's ability to learn technique. I use free software as much as possible (DSS, ImageJ) and am thus somewhat restricted to WinDoze boxes. An inexpensive box should be fine (definitely max out the RAM, tho), and just let the programs run overnight. On the other hand, if the cost of Mac software is less than the cost of a Windows computer, it would seem to make sense to buy the software?

    One other thing- I needed to get an external hard drive to give DSS enough space to stack; I suspect regardless of your box, a spacious external hard drive would be as useful as RAM.
     
  13. Sep 14, 2016 #12
    Yes, they are all tools that you would need to run from the command line. Or used within a script that you would call from the command line.

    A GUI does make life easier, no doubt about it.
     
  14. Sep 15, 2016 #13

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    It's what you pay for when you buy Microsoft Office etc.. They spend (and so does the customer, of course) a lot of money on developing interfaces with the user that are a friendly metaphor for what they (think they) are doing with their data. Nebulosity is a great example of a useful piece of software that is missing a vital layer for the user. Each new user has to learn all its quirks and be able to make sense of that enthusiastically written Manual. Fully developed software can be used from cold without reading any manual. I am getting there, with Nebulosity and it certainly seems to do the business but you never get any reassurance about what you've done. For instance, a list of the files it's dealing with doesn't seem to be available. You have to remember what you've done and get totally in tune with the mind of the programmer. At least my motivation is high enough to be achieving some small successes so I will not chuck it in the bin yet. I will probably actually buy it!!!!
     
  15. Sep 16, 2016 #14

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Absolutely. I have been getting results, of a sort, with my existing kit and I'm putting off spending any significant money until I have truly exhausted the possibilities. Tripod is pretty sturdy and my camera, although ISO 1600 is not stunning, delivers some excellent daylight pictures. My lenses are all Pentax but I do think I could benefit from a Startracker, though. At a reasonable cost of about 300GBP, it should allow my existing camera (plus lenses) to give me some pictures of some deep sky objects. I also have 1.6TB available on a recently installed internal HD.
    I started all this with the opinion that I wanted to just visual but there are so many things you can reveal with a camera. I was encouraged to see a 'smudge' where M31 should be, after using Nebulosity the other night.
    Here's a question for you, Andy (and any other experts, of course. I am not sure what sort of exposure to use and have been setting it too low, I think. Looking at the histograms, the 'black' level peak is way down and the background sky looks pretty black. Should I give it a couple of extra stops and let the software work on a very visible level of background shash, rather than setting it to what would be 'good looking' single images?
     
  16. Sep 16, 2016 #15

    Andy Resnick

    User Avatar
    Science Advisor
    Education Advisor
    2016 Award

    Glad to hear it, and looking forward to seeing what your results are!

    In terms of ISO settings and background levels, I don't think there's a clear answer. Certainly, the background level can be much higher than you think it should be.

    In my experience, I am able to get good results using images with a sky background level peaking as much as 25% according to the in-camera histogram display (DSS reports that same background level around 6%). But there is always a trade-off! The trade-off is between bright stars bleaching and obtaining light from dim stars/nebula- boost the histogram too much and the bright stars all wash out. I'm still thinking about how to image the Orion nebula: hugely bright stars surrounded by a faint nebula. Unfortunately, DSS didn't combine the subset of low ISO images (to capture the stars) and the subset of high ISO images (to get the nebula) the way I hoped, so it's back to the drawing board...

    There's also the matter of ISO noise. It's important to remember that stacking will decrease the ISO noise, but you have to stack more high-ISO images to get there. You say your camera can deliver reasonable performance at ISO 1600, so consider that your maximum- if you need more signal, either open up the lens or increase the exposure time (or both). Maybe dial the ISO down to 800 and see if you can compensate with aperture/exposure time- if you still need a little ISO boost, bump it back up to 1000 or 1250.

    And just to mention, don't forget to turn off any automatic level controls (autogain): HDR, auto-ISO, active-D-lighting, etc.
     
  17. Sep 16, 2016 #16

    davenn

    User Avatar
    Science Advisor
    Gold Member

    I agree with Andy's previous post comments

    for your K10 Pentax ... an older tech camera now days, I wouldn't be using anything over around 1200 ISO. Many of the newest cameras have much better
    signal to noise ratio with higher ISO settings and I commonly use around 3000 ISO
    Andy's suggestion of around 800 - 1000 should be acceptable for the K10 in a darker site

    Your exposure times are entirely controlled by (1) the lens focal length ... you may remember comments to others I have made in other threads about that ....
    (2) ... if using a tracker, then the f/l of the lens and the accuracy of getting the tracker polar aligned

    for your camera ....
    500 / (f/l x crop factor) = max exposure time before star trailing occurs

    eg, 500 / (50mm x 1.5) = 500 / 75 = 6.67 sec

    If you want to do longer exposures, to really get some much better images with longer exposure, you must get a tracking platform
    and then you can do individual exposures of up to several minute each and then stack a bunch of those.


    cheers
    Dave
     
  18. Sep 17, 2016 #17

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    Usual problem - money. I either buy a tracker at 300gbp or a flashy new Pentax for 600GBP+. But not in the same fiscal period (i.e. it doesn't look good, as we all know).
    Bearing in mind that people used to produce pretty good astropics with old cameras and clockwork driven trackers, I would get best bang per buck by getting a tracker.
    On the exposure matter, with a tracker, the 'blurring formula' doesn't apply . But what I was asking about was the trade off between many shorter exposures and fewer well-tracked long exposures, when you put it all in the Nebulosity melting pot. I would imagine that the Nebulosity noise reduction algorithm could be better than the general purpose algorithm in the Camera. Plus, it appears you can choose between a number of different strategies in N4.
    Interestingly, in a very few trials of the stacking process, I have found a few images that could be 'real' or just artefacts. I guess that has been a problem ever since the dawn of astrophotography. With around ten input images, I have seen rows of faint dots. I wonder whether they could be result of the 'translate and rotate' process or some slow moving object. That's something I can look into when I get better pictures. We do have a large number of aircraft around London but the total period of image gathering is longer than I would have thought an aircraft could be passing through. It certainly is entertaining to look at the blurry N4 output and to try and work out what you are actually seeing.
     
  19. Sep 17, 2016 #18

    davenn

    User Avatar
    Science Advisor
    Gold Member

    which one were you looking at for that price ? got a link ?

    yes and there are some cool homebuilt ideas out there that I have linked to in the past .....

    http://massapoag.org/astro_2/Equipment/Barndoor_Tracker/Barndoor_Tracker.html

    http://garyseronik.com/a-tracking-platform-for-astrophotography/#more-52

    http://makezine.com/projects/hinge-sky-tracker/


    there's some ideas for you
    ideal for cameras and light lenses up to ~ 100mm max ( preferably 50mm max unless you build it very sturdy)

    any serious lenses and you would really need a commercial one that will track happily with a couple of kg of payload ( cam + lens)


    Dave
     
    Last edited: Sep 17, 2016
  20. Sep 17, 2016 #19

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    The tracker I have seen (and a friend has personally recommended it) is the iOptron Skytracker v2. (Available here) Sky at Night have reviewed it and it seems to do what's needed. There is another one by Vixen which is more expensive but doesn't carry a 3kg load. It seems to be suitable for my entry level needs and I can always sell it on if I want to improve on it and spend a lot more.
    I have plenty of DIY equipment but nothing on which I could produce the sort of quality required for a tracker.
     
  21. Sep 17, 2016 #20

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member

    I just found this link. It says that why long exposures are twice as long as the shutter noise indicates, is because it takes another 'black level' picture with the shutter closed and subtracts this from the open shutter picture. Is that correct? It makes sense but it may imply that the results should be better.
    Any opinions about that?
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted



Similar Discussions: Processor power for entry level Astrophotography
  1. My Astrophotography Site (Replies: 23)

Loading...