[Bedivere]

Star photos are one area of photography that eludes me. At least as far as getting the kind of results I see other people getting. I have the basics down - how to focus, how to set exposure time, etc., but I never get that massive detail in the milky way or the crisp, bright stars against a dark background look. Some people even seem to be able to get a fair amount of foreground detail and distant city light glow into their pictures while also getting crazy amounts of detail in the Milky Way and stars. I've gotten some "OK" star pictures but none that just make me go "Wow, cool!" Here's one that I particularly like - sharply focused with an amazing amount of detail in the Milky Way:

I've also heard about "stacking" where you take several exposures then blend them together in Photoshop to bring out maximum detail. I don't really get how this works though - the stars are moving relative to the horizon. If the camera is held steady then the stars would appear to blur when the images are stacked. If the camera is mounted on a tracking device, then any foreground/earthbound elements in the image would blur. (Not to mention I find carrying a simple tripod enough of a burden on long trips, I'm not going to carry a tracker, too.) So what are your tips & tricks? How do you get the most detail into a sharply focused picture of the nighttime sky?

[NacMacFeegle]

[Jim Dockery]

I felt the same way until I finally went full frame and spent some serious $$. Star photos are quite gear intensive to get top results. A recent full frame sensor with lower noise at moderate-high ISO (2000-6400) is one key. My A7rII is one of the better cameras in this respect and I'm finding that my attempts to stack photos are way too much work compared to single shots in camera. The second key is a fast wide lens, I went with a Rokinon 14mm 2.8 which has worked quite well even though it has some distortion issues and is quite heavy. I just got a new Zeiss 18mm 2.8 specifically for stars and hope that it will be worth the extra $$ (it already will be nice since it is so much smaller and lighter). A good web site with tutorials and tech info is Lonely Speck. Another good explanation of astro-tracking, along with an incredible shot of Bryce. BTW, many of the most impressive images are composites where the stars are shot and processed differently than the foreground (even to the point of using a tracking device for the stars, then layering that over the sky in a shot with better foreground detail (sometimes shot at twilight). These are of course more art than documentary shots. Here is a stacked image I just processed today in Starylandscapestacker. It took my computer forever to process and it does a good job of finding the sky, aligning the stars, and keeping the foreground as is. Not worth the processing effort for me since my camera does such a good job on single shots.

Mt. Rainier Stars (stacked)

[Bedivere]

Thanks for the comments Jim! I had been working with a Nikon D7000 which is pretty good in low light. The only lens I had that would work is a 16-85 f3.5-5.6. f3.5 isn't bad but of course 2.8 or faster would be better. I now have a D7200 which is even better than the D7000 at controlling high ISO noise and I picked up a Tokina 11-16 f2.8 but I'm not getting very sharp results from that lens. Of course these cameras have APS-C sensors, not full-frame but I've seen some mighty good results from M4/3s and other smaller cameras so I'm not convinced that sensor size is the only issue. The D7200 outperforms full frame cameras from just a generation or so ago. The big drawback to the APS-C sensor cameras is field of view and available lens choices. Nikon makes a fantastic 20mm f2.8 that is widely available on the used market for very reasonable prices and just released a 20mm f1.8 refresh that is also reasonably priced but the FOV with that lens on a DX camera would be too narrow most of the time. Nikon makes a couple of 14mm f2.8 options but they're too expensive for me. I guess I'll just have to ask Santa Clause for a 12mm f1.8 and see what he brings down the chimney... (or maybe more realistically, a used D750 body with one of those 20mm f2.8s.)

[Bedivere]

Hmmm, there's also this: http://www.bhphotovideo.com/c/product/1003427-REG/bower_sly1620ae_16mm_f_2_0_wide_dslr.html Reviews show this to be a pretty good option for star photography, but I have to wonder whether the difference between f3.5 and f2.0 is really worth the several hundred dollars. Also, my 16-85 is a stellar performer at 16mm and probably outperforms that lens in terms of sharpness and aberration control.

[Bedivere]

Someone just posted this on Facebook. This is an absolutely amazing photo, IMO. I would love to get results like this. I wonder how they did it...?

[Jim Dockery]

That is a cool shot Will, looks like a pano. Nice detail in the mountain, but the stars on the edges are streaked. I like the headlamped hiker. If you look at the lens recommendations on Lonely Speck you'll see that going with a faster and wider lens lets you shoot longer without streaking the stars, which is key to the type of realistic shot I like, and it seems you are looking for. On APS-C that would mean a 12mm 2.8 would be great. Have you looked at this Rokinon?

[trestle]

A question from a newb, if I may: are the focal length for APS-C lenses referenced in the same way as those for micro-4/3s? So then the 12mm APS-C you just referred to would be a 35mm-format equivalent of 24mm? I hope my question makes sense, I'm trying to learn a lot in a short amount of time.

[day_hike_mike]

Just digging into this myself. APS-C sensors have a 1.5x crop factor. Micro 4/3, which I don't use so aren't familiar with, have according to Wikipedia, a 1.84 - 2.0 crop factor. On my Nikon D3300 with the APS-C sensor, my shortest focal length lens being the 18 - 55 DX kit lens is 18 MM. That works out to a 27 MM full-frame equivalent. 4/3 would be longer, multiplying the larger crop factor by the focal length of whatever lens you would have attached. 12 MM on a 4/3 could work out to a 24 MM equivalent if the factor, depending on brand I guess, was 2.0. From what I understand, to have no star movement in astrophotography, you divide 600 by the full-frame focal length to get your maximum exposure time. In my case it would be 600 / 27 = 22 seconds. I may be wrong? I want to try it out, but looks like a cloudy night tonight. May my results when we finally get clear skies be stellar!

[NacMacFeegle]

[NacMacFeegle]

Lonely Speck has a very nice Milky Way exposure calculator: http://www.lonelyspeck.com/milky-way-exposure-calculator/

[Bedivere]

Focal lengths are what they are. A 20 mm lens is a 20mm lens, period. What changes is the field of view (FOV) with different sensor sizes. This is usually called "crop factor" and has the effect of making the lens work like a longer lens on cameras with smaller sensors. 35mm equivalent sensors (often called "full frame" or "FX") are the reference for crop factor, so would have a crop factor of 1, or no crop. Nikon's APS-C sensors have a 1.5 crop factor. That means a 20mm lens will have the same FOV that a 30 mm lens would have on a full frame/35mm format camera. Canon's APS-C sensors are slightly smaller and have a 1.6 crop factor. I don't know about other manufacturer's sensors and other formats. D_H_M - I read somewhere that the calculation for maximum exposure time was 500/(focal length x crop factor). I suppose you could use 600 also, with 500 being more conservative. I've used 500 and when pixel peeping you can still see some very slight elongation of the stars, especially when shooting the Southern sky where the relative motion of the stars is the fastest.

[Bedivere]

[gb]

A great image but an awful lot of work. The guy who probably knows the most about Astro landscape is Richard Clark, who details just about every aspect of his passion in photography: http://www.clarkvision.com/articles/nightscapes/

[gb]

I do pretty good with a M4/3 Olympus EM-1 and two lenses, the 12-40mm F2.8 (24-80) and the 8mm (16mm) F1.8 FE. In general you can only shoot so long of an exposure without getting apparent star motion. The greatest motion relative to your position is probably south; the least to the north. The MW, however, in summer runs from SE to SW across the sky as the night progresses. A wider lens doesn't show this motion with a time exposure as much as does a longer focal length. This is explained by the rule of 500, which is basically the rule of 250 with M4/3. What the rule means is that (for practical purposes and using the rule of 250) divide 250 by the focal length in mm. So for my 12mm M4/3 250/12=20 seconds+-. For my 8mm M4/3 250/8=30 seconds +-. Of course, that doesn't mean the stars don't move in that amount of time; it just means that most people viewing an image will not find the amount of motion apparent. Another key aspect of astro photography is the amount of time it takes to illuminate the scene - the stars and the landscape adequately for the best exposure. This is where a blended image works best because the exposure time for the stars/sky is different from that for the landscape. (Moonlight helps landscape exposures but also tends to make the stars less visible). The amount of time in an exposure is necessary, then, for a well-exposed image, but is limited by the rule of 500 (250 for me) in the previous paragraph. That means that the only tools one really has to get proper astro/landscape exposure is by adjusting ISO, using a wider focal length, and having a wider aperture. The problem is, that as one raises the ISO, one increases the noise in the image. This noise shows up especially in the dark sky of an image. How much noise is acceptable - your choice. Some cameras are better for long exposure and high ISO noise than others. for instance, in Olympus, the sensor for the EM-5 is better than that of the EM-1 (you can Google reviews that show this noise). Since one is limited by the exposure time and the ISO, a wider aperture lets in more light without lessening the quality of an image if the lens is sharp at a wide aperture. From practical experience with my Olympus lenses, I get good exposures of the MW and the landscape on a moonless night with the 12-40 F2.8 at F2.8, 20 seconds, and an ISO between 2000 and 3200. I would not want to go above the latter ISO with the EM-1. With the 8 mm F1.8 FE I get good exposures at F1.8, where I also hold the exposure to 20 seconds (although 30 seconds would gain me 1/2 stop), and an ISO of 1000 - 2000, with 1600 normally being ideal. There is very little noise in this range of ISO's. AN UWA or my FE are also advantageous in three ways: 1) You get a wider sweep of the MW, 2) you can either shoot for a longer exposure or shoot at a lower ISO, and 3) there is so much DOF that focusing the landscape and stars simultaneously is a snap, even for foreground elements. (The disadvantage of the FE is that to keep from distorting the image, although you can use a Pannini correction, you need to keep the lens approximately level. In some landscapes one doesn't notice the distortion, in others, like a round volcano, one certainly does). I focus by walking an appropriate distance away from the tripod and shining my headlamp on some well-defined object. Focusing is easier if focus peaking is not enabled. I then shoot and review the image in magnified view, looking for the sharp outline of both near (if necessary) and distant rock or tree skylines. With the 8mm FE, I am able to focus on something as close as 7-10 feet, so that I don't even have to move. Once focused I shoot using a 12 second self timer and bracket ISO exposures. Although, I shoot mainly LSF jpegs, I also shoot RAW for this particular use. If I have any doubt about my ability to focus I also focus bracket at what appears to be the best ISO after reviewing the image. The whole process of shooting, bracketing, and waiting for those darned airplanes to fly out of the scene probably takes 2-3 hours.