Astrophotography

<p> I'm trying to shoot stars and landscapes together (like on astropics.com) Basically I want a really fast wide lens to avoid star trails. (would like to avoid stacking / stitching)<br />Some fast wide lenses that I'm looking at are the nikon 24mm f2.0, sigma 20mm f1.8 and sigma 30mm 1.4. (The new nikon 24mm 1.4 is out of my price range.) Any ideas / suggestions? I'm shooting on DX (unfortunately)<br />Something like a 10mm f1.4 would be ideal (have never seen anything like that.)</p>

<p>Thomas, do you have a particular shot or composition in mind? </p>

<p>I can get away with about 10s exposure using a 50mm/1.8 without trailing; wider will let you go longer. The general guideline I use for unguided astrophotography is ~500/focal length. </p>

<p>You'll need really dark skies if you're after the Milky Way; cranking up ISO will help but at the expense of noise. Also be prepared to work fast at light-painting the landscape. Planning is everything. </p>

<p>Before investing in an expensive lens, though, I would suggest a few practice shots with the fastest/widest lens you currently own to get a sense of what you might need in additional gear. </p>

<p>Forgot to add that my above comment is for a DX sensor. </p>

<p>I'm just starting out doing these shots. I find that 1600 ISO, 20 - 30 seconds with a 24mm/2.8 wide open gives me good results. This is with a full frame sensor. The last attempt I made was with a 17-35mm at 17mm and 3200 ISO. No light painting in the foreground, but I think it came out well. Have not attempted star trails yet. The wider the lens, the more of the sky you can include and the longer you can expose without trails. I use a regular LED flashlight when I light paint and it shows up well. I have used strobes as well.</p><div></div>

<p>Thomas: Stacking, or trying to correct star trails in FocusMagic, is really your friend if you want sharp stars. I've used my 14-24 on full frame, but it's only an incremental improvement over a DX ultrawide. A longer fast lens will show stars more brightly than a shorter fast lens, and there's no substitute for the aperture of a big telescope; the same isn't true for deep sky objects, so you might want to bear this in mind when picking your subjects. Have you considered a fish-eye? Disclaimer: I only have limited expertise here. Good luck!</p>

<p>Thomas, you might also want to consider the Samyang 35mm f/1.4 lens for this as a slightly cheaper alternative. It'll be equivalent to a 50mm lens on full-frame, which is probably about as long as you'd want to go. The fact that you'd be limited to manual focus would be no great drawback for this application. It also wouldn't be redundant should you wish to go full-frame at some time in the future.</p>

<p>Andrew. I've tried to puzzle out the paradox you're referring to; that of star brightness supposedly only being dependent on physical aperture size, while the brightness of galaxies and nebulae is governed by relative aperture. This really doesn't make sense, since galaxies are still stars, just a lot further away. Therefore the light from them ought to be even more parallel than that from more "nearby" objects, whether it can be resolved into individual points or not.</p>

<p>I suspect that this is one of those optical myths that works fine on paper with a theoretically perfect simple lens, but that in practise is found to be somewhat different when a real lens is used. Like the Cos^4 "law" that's broken by almost every wideangle lens ever made.</p>

<p>There is a limit to the smallness of spot that any real, and therefore aberrated, lens can focus a point source or star to. Therefore the light energy will be spread into that area and not into a theoretical infinitely small one. This, I believe, solves the above mentioned paradox and gives the lie to the theory that physical lens diameter is all important. Or maybe someone has a better explanation of why deep space (and earthbound) objects should obey one optical rule and near-space stellar objects another.</p>

<p>RJ - I have to say I've only been told this paradox, and not actively tried to measure it, but I've certainly seen it in several places.<br />

<br />

The explanation I've heard is that a point light source (e.g. a distant star) will resolve to a small number of sensor pixels - which admittedly is unlikely to be <i>one</i> sensor pixel - and that changing the zoom factor will not change this - the light is not spread out appreciably. The amount of light illuminating those pixels is therefore determined by the amount hitting the entrance aperture of the lens, so (for a distant source) for a constant f-stop, it should increase with the square of focal length. (For anyone trying to keep up, I'll show my working: a 50mm f/2 lens has a 50/2 = 25mm entrance aperture; a 200mm f/2 lens has a 200/2 = 100mm entrance aperture, which is four times larger linearly, and covers sixteen times the solid angle relative to the light emitted by a star.)<br />

<br />

With a larger light source, the amount of light hitting the front of the lens is spread out over more sensor pixels (or film area) as the focal length of the lens increases. A 200mm f/2 lens may capture 16x the amount of light of a 50mm f/2, but it spreads it over 16x as many pixels; this is partly why aperture in f/-terms is useful. But it only works if the longer lens actually spreads the light out.<br />

<br />

With galaxies, which are (mostly) a lot of stars from the visual perspective, the problem is that you don't have a point source - you have lots of points contributing to a finite number of pixels, and it covers an appreciable part of your image. Use a longer lens that spreads the light across multiple pixels and it will get bigger in proportion to the extra light gathering - some point sources contribute to one group of pixels, others contribute to a different group of pixels, so no single group is as well lit by all the light sources as if all the light were concentrated in once place. After all, you can consider any object as being made up of a large number of point light sources.<br />

<br />

While I've not directly measured this, I've seen a lot of stars through a 200mm f/2 that weren't there with a 14mm f/2.8 (I think the difference was more than one stop). Besides, for a telescope, f/4 is <i>fast</i>, and people have seen a lot through f/11 optics. I suspect the astonomers would have noticed if absolute aperture didn't make stars brighter. I'm still hoping to pick up a 10" dobsonian sometime...</p>

<p>Can't help you with the other lenses, but my experience with the Sigma 20mm f1.8 is that it's hopelessly soft below f5.6, and essentially useless for what you're trying to do...</p>

<p>Joseph: thanks, that is good info, stopping down to f5.6 is not an option.<br>

Andrew: I have considered (and plan on using) a fisheye. I have the nikon 10.5mm 2.8 which I got some nice shots on Haleakala with (if I had to do it again I would have used higher ISO.) I think focus magic will only work if shooting with a long lens and having the star trails go in a straight line (with the wide lenses all the lines are going to be curved in varying amounts.)<br>

Joe, I'm looking at that samyang 35 1.4, and now strongly considering the sigma 30mm 1.4. Having that fast of a lens will help and if I decide to do some stitching I'll still have the speed.</p>

<p>Michael, thanks for reminding me of that formula, I had seen it before but forgot about it. That will definitely come in handy as I try a few different lenses out. ( I basically want to have whatever interesting foreground object on the ground beneath a nice view of static stars in the milkyway.)</p>

<p>Thomas: I second (from very limited experience) Joseph's comment about the 20mm Sigma - I tried one in a store and it was pretty horrendous. Re. Focus Magic, it's true that it can only fix straight lines - but if you process sections of the image separately, and if the exposure is short enough that the trails are approximately linear (if not in the same direction in different parts of the image), it ought to be able to help. That said, I've only meant to try this, not actually done so... Good luck, and I'm envious of the opportunity. I must go and visit my sister in Wales (or try to work out how to fit a sodium filter to a 14-24).</p>

<p>You might want to check out my set of photos that I have on Flickr. They're not very professional, but you might find them helpful since I give the exposure times. That was before I learned about the 500/focal length formula.<br>

I basically used an 85mm, 24mm, and a 16mm fisheye.<br>

<a href="http://www.flickr.com/photos/nathantw/sets/72157619526692330/">http://www.flickr.com/photos/nathantw/sets/72157619526692330/</a></p>