In May of 2015 Nikon provided me with their brand new D810A to give it a workout for landscape astrophotography. Check out my article on Nikon’s Image Chaser website.
The Short Answer
In my experience the D810A matches the high ISO performance of the D750. The D750 has a 24MP sensor and until the D810A came out the D750 was the cleanest high megapixel DSLR, about a stop better than the D810. But the D810A, like the D800/D800E/D810, has a 36MP sensor and appears to have the same high ISO performance of the 24MP D750. With some other nice features, the D810A is an amazing landscape astrophotography camera.
The Long Answer
Nikon 14-24mm f/2.8 lens
Star stacked sky from 10 x 10s f/2.8 ISO 12800
Foreground multiple exposures at ISO 1600 f/2.8 15 minutes
ISO 12800 Comparison
For shooting landscape astrophotography when you’re not using a star tracker or star stacking (see Star Stacking section below) for pinpoint stars and lower noise, you want to use the highest ISO possible with your camera that produces a useful image when taking your sky exposures. (I take separate exposures for the sky and foreground at different ISOs and blend them to get the best depth of field and a cleaner foreground.)
Generally speaking, a more brightly exposed shot with a high ISO will be cleaner than an underexposed shot with a lower ISO, even if the difference is only 1 exposure stop (half the brightness). Boosting the exposure in post of the lower ISO image to match the higher ISO image may show a similar result, but upon close inspection there is often a distinct difference, with the higher ISO shot having less noise. This is a result of an increased signal in the higher ISO shots, the noise floor is closer to the exposure in the lower ISO shots and thus you end up with more visible noise after boosting the exposure in post.
This result will depend on your camera. Test various ISOs for your use.
For my tests with the D810A, D810, and D750, I used my Nikon 14-24mm f/2.8 lens @ 14mm and f/2.8 with a 25 second shutter time for all ISO test shots. You need to chose a low enough shutter time to limit star trails, and when not star stacking I prefer to use 25 seconds when shooting at 14mm. I find this results in acceptable star trails while still getting a bright enough exposure.
I chose to use ISO 12800 for my primary tests because it is the highest native ISO of all 3 cameras. And since, in my experience, a 25 second ISO 12800 exposure will be cleaner than a dark 25 second ISO 3200 exposure, I have been using ISO 12800 for most of my sky shots while testing the D810A. I even used ISO 12800 for panoramas without using a lower ISO for the foreground and with enough balancing of noise reduction and sharpening the foreground can work out amazingly well.
Here are the ISO 12800 shots from each camera.
Note that the raw NEF files for all shots in this article were processed with Capture NX-D 1.2.1. As of this writing Capture NX-D is the only raw editor capable of reading D810A files. My usual raw editor is Adobe Lightroom.
|Nikon D810A, Nikon 14-24mm f/2.8 lens, ISO 12800, 25 seconds, 14mm, f/2.8|
|Nikon D810, Nikon 14-24mm f/2.8 lens, ISO 12800, 25 seconds, 14mm, f/2.8|
|Nikon D750, Nikon 14-24mm f/2.8 lens, ISO 12800, 25 seconds, 14mm, f/2.8|
Now lets take a look at 100% crops from each camera. Note that there has been NO noise reduction whatsoever applied to these images. Noise reduction in Capture NX-D was not used. Realistically, you would at least apply some color noise reduction. If you use Lightroom, color noise reduction is always enabled by default. For a fun experiment, go find a high ISO, say 6400 or so, raw image in Lightroom, zoom into 100%, and turn the color noise reduction slider to 0 and watch how splotchy the image becomes.
|Nikon D810A, 100% crop, no noise reduction|
|Nikon D810, 100% crop, no noise reduction|
|Nikon D750, 100% crop, no noise reduction|
As you can see, the D810A and D750 are pretty close in noise, although the D810A has more color in the sky. The D810 clearly has more noise. Although arguably the noise difference isn’t that dramatic and can certainly be cleaned up in post. And remember that if you’re star stacking it doesn’t matter.
Now how about the darker foreground areas? For these examples I applied color noise reduction to TIFF files in Lightroom. The TIFF files were exported from Capture NX-D. Capture NX-D has color noise reduction but Lightroom’s color noise reduction appeared to work much better. I used a value of 100 in the Color noise reduction slighter in the Detail panel of Lightroom for each image. This a realistic value that I might use for such noisy foregrounds. There is no point in comparing the images without any color noise reduction as that is an unrealistic application. I may not use such aggressive color noise reduction on the entire shot and only apply that to the foreground section if I’m not using separate foreground exposures.
|Nikon D810A, ISO 12800, 100% crop, Lightroom color noise reduction of 100|
|Nikon D810, ISO 12800, 100% crop, Lightroom color noise reduction of 100|
|Nikon D750, ISO 12800, 100% crop, Lightroom color noise reduction of 100|
For whatever reason the D750 exposure came out darker than the others. I’m not sure why, perhaps it is an issue with the D750 itself, or Capture NX-D. I brightened the D750 exposure by 1/2 a stop to better match the brightness of the other exposures. Without brightening it the color noise is far less visible.
They’re all pretty darn noisy, but the D810 is certainly much worse off than the D810A and D750. This difference alone is a huge boost for using the D810A or D750 for when you don’t have time to take additional shots for the foreground at a lower ISO, or low enough ISO. I like to use ISO 1600 or lower for foreground shots when I can, but in a pinch I know I can rely on the performance of the D810A or D750 to get by with a higher ISO and more clean up in post.
While not specific to the D810A, I feel that it is important to show how much of a difference in noise you can get by using the star stacking technique. This applies to any camera and can take images from a very noisy camera and still produce a clean sky.
Star stacking is a technique that involves taking multiple very high ISO short exposures to capture pinpoint stars without any visible trails, and then stack, align, and average them in software to vastly reduce the noise. I use Starry Landscape Stacker for Mac, available in the Mac App Store for just a few dollars. You can do it in Photoshop, but Starry Landscape Staker makes the process much easier. Deep Sky Stacker is a Windows program that might work for this but I’m not sure how it will handle the alignment process with foreground elements. Deep Sky Stacker and similar programs will align multiple images based on the stars but Starry Landscape Stacker is designed to handle the foreground and only align based on the stars.
Using 14mm on full frame, I generally take 10 exposures at 10 seconds each at ISO 12800 or whatever the highest native ISO is on the camera I’m using. However on many cameras you may need to use a lower ISO to avoid excessive magenta color noise on the edges of the frame. On the D810 you might want to use ISO 6400 instead. On my D800E I had to use ISO 5000 or 4000. This color noise is due to both under exposure in the corners from vignetting of the lens and sensor noise from heat and other electronics in the camera. On the D810A I’ve been able to get away with ISO 12800.
Here is an example:
|Nikon D810A, ISO 12800, 100% crop, 10 seconds, f/2.8|
|Star stacked result using 10 exposures of 10 seconds each at ISO 12800.|
Now you can see why I use this technique whenever I can!
Effects of the IR Filter
The D810A is really designed with deep space astrophotography in mind due to the special IR filter that is optimized to capture the H-alpha narrowband infrared light emissions from nebulae. This results in red nebulae that really pop in the image, but also has the side effect of causing other light sources to shift. Daytime photos can sometimes result in a red cast. Light pollution at night can take on a yellow glow instead of orange.
Here is an example of how the IR filter pops red nebulae:
|Nikon D810A, ISO 3200, Nikon 50mm f/1.4G lens, f/2, 30 seconds, iOptron SkyTracker (camera mount for star tracking)|
|Nikon D810, ISO 3200, Nikon 50mm f1/4G lens, f/2, 30 seconds, iOptron SkyTracker (camera mount for star tracking)|
Even without zooming in the difference should be noticeable. The red nebulae are more noticeable and the Milky Way takes on more color with the D810A. But lets take a closer look:
|Nikon D810A, 100% crop|
|Nikon D810, 100% crop|
IR and Light Pollution
Now let’s look at the difference with light pollution. This difference can also be seen in the uncropped images at the start of this article, look at the glow in the lower left of the frame and on the D810A it looks yellow vs orange on the others.
|Nikon D810A, ISO 3200, Nikon 50mm f/1.4G lens, f/2.8, 30 seconds|
|Nikon D810, ISO 3200, Nikon 50mm f/1.4G lens, f/2.8, 30 seconds|
Notice how the light pollution glow in the D810A image looks yellow instead of orange.
IR and a Lighthouse
|Nikon D810A, ISO 12800, 10 seconds, 24mm, f/2.8|
|Nikon D810, ISO 12800, 10 seconds, 24mm, f/2.8|
IR and Daytime
The IR filter in the D810A really means that the camera is best suited for the dark night sky, and there will be a red cast during daytime shots, but how noticeable or problematic it is will depend on the situation. Here is an example.
|Nikon D810A, ISO 200, 24mm, 1/50s, f/11|
|Nikon D810, ISO 200, 24mm, 1/50s, f/11|
The white balance on both of these shots was set to Direct Sunlight in Capture NX-D.
Notice in the D810A photo how there is a reddish tint to the rocks and trees in the foreground and background, the water is a warmer blue, and the hazy distant horizon also has a reddish tint.
M* Manual Mode – Exposures Longer Than 30s Built In!
One of the most welcome D810A features for long exposure photography is the ability to choose exposures longer than 30 seconds in camera without the need for a remote timer. Put the D810A in M* manual mode and after 30s you can choose 60, 90, 120, 240, 300, 600, or 900 seconds. I really wish there were more options but at least it’s a huge step in the right direction.
If you enable Exposure Delay mode and very gently hit the shutter button on the camera you can try to get away without using any remote at all as long as you don’t need other long exposure times. While testing out the D810A in Acadia I shot for almost the entire week and a half without the use of a remote.
|M* Manual mode on the D810A with a 240 second exposure selected.|
Capture NX-D Astro Noise Reduction
Capture NX-D, as of version 1.2.1 (released in time with the D810A) has a new Astro Noise Reduction checkbox in the Noise Reduction panel. Enabling this will remove the majority of hot pixels from an image. It’s not 100% perfect though, but it works pretty darn well. On the other hand Phase One’s Capture One raw converter has VERY good hot pixel noise reduction. It has some built in by default, and gets better with enabling color noise reduction, and then using the Single Pixel slider can get rid of all or almost all hot pixels. There’s also PixelFixer (Windows program) for doing the same thing with or without supplying your own dark frames. These are all very useful tools, and if you don’t have time for Long Exposure Noise Reduction in camera you can use one of these tools to fix the hot pixels in post, or at least get most of them and spot clean the rest yourself.
And let’s not forget the Dust & Scratches filter in Photoshop. That can produce excellent results with care and doesn’t require a big change in your raw editing workflow.
I also have outdoor workshops available: