Click here if you are having trouble logging into the forums
Privacy Policy |
Please read our Terms
of Service | Signup and
Troubleshooting FAQ | Problems? PM a Red or a Green Gu.... uh, User
diggy
sage
Reged: 05/09/04
Posts: 288
Loc: Vancouver Island, Canada
|
|
I've been stumbling on the black art of focusing a DSLR at prime focus without the recommended computer connected tools. One question is how big would a star be in pixels, if the image is "ideally" focused. I realize the question is probably over-simplified: I'm sure it depends on a bunch of factors. So, for an example:
I shot the double cluster (12 subs 25" ISO 400) with a Nikon D70s (3008x2000 pixels) through a 110mm refractor with a 0.8 focal reducer/field flattener. The "largest" star in any of the raw images measured 12 pixels across; the smallest about 4 pixels wide. After processing through DSS the largest was reduced to 9 pixels. So, is there any info I can gather from that about focus: should stars be smaller?
Thanks in advance for any guidance here.
Diggy
-------------------- WO 110mm Megrez doublet on an HEQ-5 mount
Nikon 10x50 binocs
"Beyond here be there"
|
imjeffp
Senior Space Cadet
   
Reged: 09/30/03
Posts: 4427
Loc: Cedar Park, Texas
|
|
Hmmm... I suppose you could do the math and figure the Airy Disc size for your scope, and the arc-seconds/pixel of your imager and combine them.
-------------------- Blog
ST80 • AT80EDT/LXD650
ETX-90/DS-2000 • 10" LX200 Classic ("The Quarter-Meter Telescope at the Heritage Park Observatory")
SPC900NC • DMK21AF04 • Digital Rebel XT
|
NeoDinian
Experienced Postmaster
   
Reged: 10/05/05
Posts: 12070
Loc: Rockford Illinois
|
|
Quote:
I've been stumbling on the black art of focusing a DSLR at prime focus without the recommended computer connected tools. One question is how big would a star be in pixels, if the image is "ideally" focused.
A star is ALWAYS a "Pin-Source" of light. So in theory, the "Pixel size" of a star "SHOULD" be 1 pixel... Problem is, brighter stars will make adjacent pixels glow as well. (Not to mention that pics with 1-pixel stars would look kinda funny...) This is why we see stars of different sizes on the images. But in reality, the star itself is only the center pixel... The rest of what we see is the "Glow" from the exposure...
-------------------- Neo... (Jeff)
-=-=-=-=-=-=-
10" LX200-GPS/SMT UHTC "Draco"
Rockford, Il.
NeoDinian's Eye on the Sky!
Coming soon:
Edited by NeoDinian (08/09/08 11:37 PM)
|
gavinm
professor emeritus
Reged: 08/26/05
Posts: 559
Loc: Auckland New Zealand
|
|
You would need to know your pixel scale and local seeing, and use a fairly short exposure. I would suggest you are making life difficult by not using any number of software tools.
A quick mental calculate would get about 1-2 arcsec/pix, so smallest star size of 4 pixels equates to about 6-8 arcsec - largest 20+ arcsec (either really out of focus or bad seeing) - but a start
-------------------- Gavin
Mt Albert Grammar School Observatory
Auckland, New Zealand
http://www.mags.school.nz/astronomy/index.html
12" LX200R w/ UHTC
SBIG ST7-XME + CFW10
Skywatcher Equinox ED80 Pro (ADM dovetail)
+ other stuff
|
gavinm
professor emeritus
Reged: 08/26/05
Posts: 559
Loc: Auckland New Zealand
|
|
Make that about 3 arcsec/pix, so star sizes between 12 and 36 arcsec.
-------------------- Gavin
Mt Albert Grammar School Observatory
Auckland, New Zealand
http://www.mags.school.nz/astronomy/index.html
12" LX200R w/ UHTC
SBIG ST7-XME + CFW10
Skywatcher Equinox ED80 Pro (ADM dovetail)
+ other stuff
|
Wouter D'hoye
sage
Reged: 06/27/03
Posts: 370
Loc: Belgium
|
|
Hi,
How large a star appears in an image depends on quite some factors. First of all the optics. Every optical design shows stars differently. And the size and shape can vary depending where the star is located in the field of view. (Think about coma, astigmatism,...) Many instruments try to get this therorethical starsize as small as possible. Numbers between 5 and 20microns are quite common. Then there is seeing. if a star dances a little in he sky it gets smeared in the image, looking bigger. very bright stars will also activate nearby pixels making it glow more than less bright stars.
Wouter.
-------------------- 4" f/8.8 Vixen Fluorite on GP-DX
SBIG ST-7 + CFW8a + Baader CLRGB-filters
|
ccs_hello
Carpal Tunnel
   
Reged: 07/03/04
Posts: 2805
|
|
Diggy,
Since you use a DSLR, there are two additional factors:
1. It has an anti-alias filter thus will spread incoming light, even it's true point light source.
2. It uses one-shot primary-color Bayer (RGBG) CCD, thus it's 4 phototsites forming a decodable "white" point-light-source.
Clear Skies!
ccs_hello
|
cam1936
sage
Reged: 08/01/08
Posts: 242
Loc: Alberta, Canada
|
|
Every star, as stated above, should technically appear as one pixel. This is impossible due to the effects of the atmosphere and telescope optics ect. On a side note only the largest telescopes on earth can resolve the size of a star directly and even then it's only a handful of stars that are large and realativly close, like betelgeuse.
-------------------- C8-SGT
Nikon 10x50 Action Ex
Canon 350D
Manfrotto tripod
Various EPs and accesories
|
jay52
Carpal Tunnel
   
Reged: 09/26/04
Posts: 2280
|
|
Diggy:
Obviously, when focusing, you want stars to become as small as you can. Certain software, like MaxIm, has a way of measuring the average star width using FWHM, which is a measure of the width of the star at the point where the star is half its peak-center brightness. This is a consistent way to measure star size regardless of the stars brightness independent of the number of pixels it covers.
So, I would echo another posted above...look into some tools that do this for you. If you have a robotic focuser, then the freeware FocusMax is all you need. Likewise, if using a DSLR, then Images Plus also has those capabilities built in. Even if you focus manually, you can use it to judge your focus images.
But if you continue to focus manually, realize that as stars get smaller, fainter stars begin to pop up...and they will very often be ONLY one pixel in size. So, in focusing, look for the faint ones, not the bright ones. If you focus and a star pops in, and then out, of view, then you've gone too far.
-------------------- jay
www.allaboutastro.com
|
Arkalius
professor emeritus
Reged: 08/03/06
Posts: 668
Loc: Orange County, CA, USA
|
|
Because we're using aperture to view these things, the size of the star itself really doesn't matter (it's far smaller than what the optics are capable of magnifying). Instead, the size of the image of the star is the most important thing. With an obstruction-free perfectly round aperture, a point source like a star appears as a series of light rings expanding away from the center. The first couple are the brightest and subsequent rings are very faint. How large a perfectly focussed star will be in an image depends on how bright it is and how long you expose for. The brighter and longer the exposure, the more of these diffraction rings you will image, and the larger the star will become.
The fainter stars will only show the first diffraction ring typically in a shorter exposure. How big this first ring is depends on the size of the optic and the obstruction, if any. All other things being equal (including image scale), larger optics will produce smaller stars.
Let's say you're imaging with a scope with 5 inches of unobstructed aperture and 1250mm of focal length. This is an f/10 scope. You take a 20 second image and note the size of the stars. Now, lets say you move the same camera to a scope with 10" of unobstructed aperture, and 1250mm of focal length. You'll have the same image scale, but now you're shooting at f/5, two stops faster. So, we reduce our exposure time to 5 seconds to compensate for having 4 times the light. The fainter stars that showed only the first diffraction ring in the previous photo will appear to be half the size in this photo, because we have twice the aperture. When you add an obstruction it complicates things because the obstruction adds it's own diffraction rings that repeat at a different interval than the others causing different interference pattersn. It typically has the effect of brightening some of the outer diffraction rings of a star image which will in effect make the star appear larger.
jay52 mentioned the FWHM (Full Width at Half Maximum) method of measuring a star's size. There's another method called Half-Flux Diameter which measures the circular area of the star image that contains half of all the light energy, and it is typically smaller than the FWHM for the same star. FWHM is the preferred method typically.
-------------------- -Arkalius
11" Celestron SCT on Orion Atlas EQ-G
8" Zhumell Dobsonian Reflector
Proud member of the Orange County Astronomers
|
Jared
Carpal Tunnel
   
Reged: 10/11/05
Posts: 1871
Loc: Piedmont, California, U.S.
|
|
Quote:
I've been stumbling on the black art of focusing a DSLR at prime focus without the recommended computer connected tools. One question is how big would a star be in pixels, if the image is "ideally" focused. I realize the question is probably over-simplified: I'm sure it depends on a bunch of factors. So, for an example:
I shot the double cluster (12 subs 25" ISO 400) with a Nikon D70s (3008x2000 pixels) through a 110mm refractor with a 0.8 focal reducer/field flattener. The "largest" star in any of the raw images measured 12 pixels across; the smallest about 4 pixels wide. After processing through DSS the largest was reduced to 9 pixels. So, is there any info I can gather from that about focus: should stars be smaller?
Thanks in advance for any guidance here.
Diggy
The short answer to your question is, "Yes, stars should be smaller."
The longer answer is a bit more complicated. Your D70s has 8 micron pixels. That is, each pixel is 8 millionths of a meter across. Just divide the 24mm chip width by the 3,000 pixels to get this value.
If you have typical seeing conditions--somewhere between two and three arc seconds across--those 8 micron pixels are going to be somewhat undersampled using your refractor and focal reducer. You didn't mention the focal ratio of your scope, but if it's around f/7 you are looking at a little over 3 arc seconds per pixel. For optimum sampling with 2-3 arc second seeing conditions you would want more like 1 to 1.5 arc seconds per pixel (some would argue even a bit lower). As a result, your resolution on most nights is going to be limited by your camera, not by the seeing conditions or the telescope. That's not bad, by the way. It's common place in wide field astrophotography where resolution is compromised in order to obtain a wider field of view. This is spatial resolution we are talking about, not related to the apparent sharpness of the image.
So where am I going with all this? Since you don't have tools to measure FWHM (full width/half maximum) values, you will need to judge just from looking at the stars themselves. Since your system is undersampled for typical seeing conditions, it should be possible to focus to the point that faint stars occupy only a single or perhaps a couple of pixels. That's what you should be shooting for in your focusing. As Jay mentioned, a good way to judge this is by how bright the faintest stars appear to be. When minute adjustments are causing faint stars to appear or disappear, you are right near the point of optimum focus.
If faint, small stars are 4 pixels across right now you can probably do a bit better with more careful focusing. I would recommend some software tools to help you evaluate your results. It makes focusing much easier.
-------------------- - Jared Willson
- Fluorostar FLT-110 w/ TEC optics
- Vixen VC200L
- Astro-Physics Mach1 GTO
- Stellarvue SV80S
- Takahashi Teegul SP Mount
- STL-11000
|
diggy
sage
Reged: 05/09/04
Posts: 288
Loc: Vancouver Island, Canada
|
|
Hey, thanks Jared for the clear and detailed answer. I'll scratch my head and puzzle out all these variables eventually. My scope, by the way is f/5.9, but used with a 0.8 focal reducer. So, if I understand it, I'm probably getting more like 4 or 5 arc seconds per pixel, which means even more undersampling with the Nikon D70. I'm also thinking that exposure time would affect this whole situation. Let's say I'm imaging clusters (stars only). If I over-expose, I'm going to flood adjacent pixels. So, assuming good focus, my next mission would be to optimize (i.e. minimize) FWHM for my seeing conditions. The tip about faintest stars equaling one pixel is a good one. Yikes, what a lot of stuff to juggle 
Thanks again, Diggy
-------------------- WO 110mm Megrez doublet on an HEQ-5 mount
Nikon 10x50 binocs
"Beyond here be there"
|
groz
professor emeritus
   
Reged: 03/14/07
Posts: 541
Loc: Duncan, BC
|
|
Quote:
So, is there any info I can gather from that about focus: should stars be smaller?
Thanks in advance for any guidance here.
Diggy
I struggled with focus early on, tried a few different software packages, all of which focus on metrics like FWHM etc, and never had a 'good' result from it. Then one day I stumbled on a different method, that actually works really well for me.
Like you, I am on the island, and, we have a great view of the northern sky from here. I use this technique with both the Nikon D-50 and the Canon 300D. I always start by aiming at the alcor/mizar pair. Looking thru the viewfinder, it's relatively simple to get focus in the ballpark on this widely spaced pair. Once I'm close thru the viewfinder, I switch to the 'shoot/download/analyze' method. I'm using motor focus on the scope, so I can make very fine adjustments, semi-repeatable.
For the second stage of the focus, I completely ignore star brightness, and any metrics like FWHM etc, but instead concentrate on mizar, it's a double. Set the camera to jpg fine for faster downloads, and use 2 second exposures, then adjust focus to cleanly split the mizar double. That gets us 'real close', and when that double is cleanly split, you will see another star (mag 7) starting to show up between alcor and mizar. You may have to bump exposure time up a little to see it. Once that one shows in the frame, use it to further fine tune focus, maximizing its brightness. When you get 'really close', it may start saturating and you need to shorten exposure again, back down to 1 or two seconds.
Once I've got that nailed, I slew onto my final target, and take a 10 or 20 second exposure to validate my framing. To finalize focus on my final framed position, I like to use at least 15 second exposures, home in on some very dim background stars that _just_ show up in the frame, then make fine focus adjustments to once again maximize the brightness on those dim stars. Using specific numbers with the 127mm mak and the 300D, when I'm 'close' to focus, I will be able to clearly distinguish mag 9 or 10 stars on a 15 second exposure. When focus is right on the money, mag 11 and 12 stars will start popping out, which were not visible on the earlier shots.
Using the nikon camera, it's sometimes tough to find software that actually works and 'helps', but, there is stuff out there that can be quite useful. Here is a technique you can try, and you will be quite surprised at how well it works. Start Deep Sky Stacker Live, then as you take focus shots, download them immediately into the folder DSS Live is monitoring. Watch the statistics on the frames as they get absorbed by the program, ignore FWHM initially, and just look at star count. Using shorter exposures (10 to 30 seconds), the star count will start to jump up dramatically as you get very close to focus. I use jpg with the canon 300 for focussing because it's usb 1 and VERY slow to download (62 seconds for a raw image), but the nikon we just use raw all the time, it's usb2 and downloads are snappy.
Where are you located on the island ? We are in Duncan, and will have a new mount to try out just as soon as this weather system finishes moving through. If we got together for an evening with a clear sky, I think we can get you repeatably focussing that system 'dead on' with just a couple hours of 'playing'. Are you going to be at the Rascals star party end of the month on the malahat ? We will be there, easy to spot us, his and hers astrophotogrpahy setups. His = EQ6-Pro with skywatcher telescopes mounted sid by side, Hers = HEQ5-Pro with a Williams Optics Zenithstar 80 hosting a Nikon D-50+300mm lens piggyback. A year ago we couldn't get anything focussed repeatably, today, we can set up, polar align, and focus both systems in about half an hour. Up until now, 'hers' has been on the NEQ3, but, the HEQ5 should arrive today or tomorrow And yes, I'll plead guilty to inviting in 2 weeks of cloudy weather by purchasing new gear, at least I had the foresight to time it on a full moon cycle
|
jay52
Carpal Tunnel
   
Reged: 09/26/04
Posts: 2280
|
|
BTW, if you use FWHM, be sure you measure stars that are less than half saturated...not peaked at 65535. This assure that your response remain linear and will give you consistent measurement across the entire frame.
-------------------- jay
www.allaboutastro.com
|
diggy
sage
Reged: 05/09/04
Posts: 288
Loc: Vancouver Island, Canada
|
|
Hey, I think we met at the Malahat star party earlier this summer...I was the guy next to you with the new Megrez 110: and we got rained out. I remember both your setups. Yes, I plan to be at the Rascals star party. So we can get some "focus time" in there. Thanks in advance!
Diggy
-------------------- WO 110mm Megrez doublet on an HEQ-5 mount
Nikon 10x50 binocs
"Beyond here be there"
|
|
8 registered and 11 anonymous users are browsing this forum.
Moderator: Charlie Hein, knuklhdastnmr
Print Thread
|
Forum Permissions
You cannot start new topics
You cannot reply to topics
HTML is disabled
UBBCode is enabled
|
Thread views: 375
|
|
|
|
|
|
|