Moon Photography: Stellarium, Focal Length, and Exposure Times.

On Apr. 15, there will be a total lunar eclipse visible from almost the entire North American continent. The next total lunar eclipse visible from the east coast will not be until 2015, and the next total lunar eclipse visible from the entire North American continent will not be until 2019, so this is a rare opportunity to get some really unusual photographs.

Here is the field of view one can expect from a few common configurations:

D800E @ 200mm with 2X TC

D7000 @ 200mm with 2X TC

D7000 @ 200mm

D800E @ 200mm

D7000 @ 50mm

D800E @ 50mm

How to make simulations like these for any camera:

(Or skip to exposure time section)

Stellarium is an open source planetarium software that is capable of simulating the sky and celestial bodies at any time as seen anywhere on earth (and from most other known planets and moons). It's a good way to know exactly when and where you should be looking based on your exact location. If you are photographing through a telescope, it also can drive many telescopes to track the moon and other celestial objects.

General Stellarium Setup

Location selection is the top option on the leftmost menu. You can search for a city or enter precise GPS coordinates if you have them, or just click around the map.

The second option on the left toolbar is the date and time, which you should set to about 1:00 on April 15th. On the right side of the bottom toolbar the fastforward, rewind, normal, and current time buttons allow you to travel back and forth in time through the entire event.

Once you have the location, date, and time selected, Stellarium will be displaying a view of the night sky that allows you to zoom and pan around and click on celestial bodies for information.

Camera and Lens Simulation (Oculars plugin) 

However, we're interested in simulating the view through a specific lens onto a specific camera sensor. To do that, we want to use the Oculars plugin. Oculars may be enabled automatically, but if it isn't, it can be enabled under Configuration Window > Plugins > Oculars. Be sure that the "Load at startup" option is selected.

The Oculars configuration window is where you can specify a camera and lens combination. Here's an overview of the important parts of the settings panel:

General	As far as this guide is concerned, check all three boxes on top and enable the on-screen interface.
Eyepieces	This is where eyepiece information is specified; it can be ignored for astrophotography because we are targeting a sensor as the final element.
Lenses	This is where you should specify a teleconverter, if applicable, not the lens you are using.
Sensors	This accepts information about the exact sensor you're using. Stellarium uses this sensor information to calculate crop factor and other information, but for a quick-and-dirty view, the pixel seze can be omitted (just put in 4.8).
Telescopes	This is where information about your lens goes. Diameter can be omitted (set to 80) if you just want a field of view estimate.
About	This one explains itself. That's the point.

Now you should have four buttons in the upper right corner: telescope view, sensor view, scope view, and settings. Selecting sensor view (the rectangle) draws the red rectangle on the screen and opens up a menu where you can select the camera and lens setup you entered earlier. For example, the image on the left shows the simulated view of the setup I plan on using on the 15th.

Exposure Time

Again, I'll offer a cheat-sheet before my derivation:

Nikon D700, APS-C 16.2 mp, with no motion blur:
200mm lens: 1/4s or faster
400mm lens: 1/6s or faster
1250mm lens: 1/20s or faster

Nikon D7100, APS-C 24.1 mp, with no motion blur:
200mm lens: 1/4s or faster
400mm lens: 1/8s or faster
1250mm lens: 1/25s or faster

Nikon D800E, Full frame 36.3 mp, with no motion blur:
200mm lens: 1/4s or faster
400mm lens: 1/8s or faster
1250mm lens: 1/25s or faster

The D7100 and D800E have the same speeds because even though the D800 has more resolution, its pixels are spread out over a wider area then the D7100, and it just happens to round out roughly equal.

In general astrophotography, some people use the "Rule of 600" to estimate how long of an exposure time we can use before the stars start to visibly exhibit motion blur. It states that 600 divided by the 35mm equivalent of the focal length of the lens gives the exposure time in seconds of an acceptably sharp image. For example, the 200mm lens with a 200mm teleconverter on my D7000 is 600mm equivalent, and 600 / 600 = 1, so 1 second is approximately the longest amount of time I should expose the stars. That 400mm lens on my APS-C camera has an angle of view of about 3.34° (2arctan((35)/(2*(1.5*400)))), projected onto 4,928 horizontal pixels, so each degree of view is projected over 1,475 pixels. Therefore, assuming the worst-case scenario is that the fastest stars will appear to move at 0.0042 arc minutes per second (360° over 24 hours), 1 second is 0.0042' of movement, which equals a blur of 6.1 px.

The moon moves 13° more per night then the background stars, so it would have 6.3 pixels of motion blur with the same lens and one second exposure. Because the moon is the entire focus of the image, I think 6.3 px is way too much blur. To reduce the moon-motion blur to three pixels, we would need to expose for no more than 1/2 second, and to remove motion blur entirely, no more than 1/6 seconds.

This means that acceptable exposure times go down drastically with the length of the lens; my 1250mm telescope will have three pixels or more of motion blur at 1/6 second, and it will have to be faster than 1/20s in order to actually freeze the moon.

Keep in mind that my D7000 does not have an incredibly high resolution sensor, and to freeze motion on a higher resolution camera would require faster speeds. The D7100 has 24.1 mp across the same image sensor area, so through the 400mm lens it would require a shutter speed of 1/8 to fully freeze the moon, or 1/25 through the telescope. Achieving proper exposure with these shutter speeds isn't an issue with a brightly lit moon, but when it passes into the sun's shadow these limits may come into play.

Keep in mind that taking photos at faster shutter speeds than these will not improve sharpness, (assuming that the camera is on a tripod and triggered by a remote shutter) so it's better to use these speeds with as low of an ISO as possible instead of boosting ISO to increase shutter speed.

Sharpness: Aperture and Diffraction

18-55mm @ 18mm f22 15s ISO 100

This is the article I wish I'd read before traveling to France and taking pictures like the one on the right. In the full resolution image here, fuzziness caused by diffraction is clearly noticeable.

Depth of field and Diffraction

Changing the aperture of a camera is the best way to change the depth of field. Larger apertures produce images with smooth, dreamy qualities that have a very depth of field, while smaller apertures produce images with a much larger depth of field. This may lead some photographers to think that the smaller the aperture is, the sharper the image will be, but that is not true.

When light rays travel through a small hole they bend and interfere with each other, an effect known as diffraction. Unfortunately, this creates a difficult tradeoff for photographers, because stopping down the lens to increase depth of field also increases diffraction. The result is a photo that has a large depth of field but is quite fuzzy.

This issue gets bigger at larger formats, because increasing the size of the sensor (or film) decreases the depth of field. This is why medium format cameras have aperture settings far smaller then f22 (the most common minimum aperture on modern lenses built for digital APS-C and FX sensors). lenses for medium format often have minimum apertures of f45.

The Practical Test

I set up a still life and shot it with three different cameras to demonstrate the effects of diffraction and pixel size. The D50 has 6.1 MP sensor, the D7000 has a 16.2 MP sensor, and the D7100 has a 24.1 MP sensor. The reason I chose to demonstrate with three cameras at three different resolutions but the same sensor size (Nikon DX, also known as APS-C), is because the pixel size on the sensor matters. A camera with larger physical pixels is less effected by diffraction than a camera with smaller pixels packed more tightly together because the ratio between the pixel size and the diffraction pattern (called an airy disk) size is smaller. This Cambridge in Color page goes into more detail. The best way to demonstrate this would be by comparing two cameras with different sensor sizes but the same resolution, like the Nikon D7100 and Nikon D600, but I don't have a D600 available.

Cameras:

50mm f1.4 @ f8 1/60 ISO 100

50mm f1.4 @ f8 1/60 ISO 100

50mm f1.4 @ f8 1/60 ISO 100

I didn't end up posting the D7000 images on this page, but I'll link to the full images at the end.

Test shot:

I set up this still life with an old book because it has a very high level of detail—the D7000 at f8 can see individual printing dots—and can be parallel to the focal plane to try and remove some of the effects of the depth of field changes. All test shots were shot on a tripod with a tether or remote shutter and are unprocessed raw images straight from the camera that have been aligned and white balanced. All test images were shot through the same 50mm f1.8 lens at ISO 100, except for the D50 which only goes down to ISO 200.

D7100 50mm f1.8 @ f8 2.5s ISO 100

Starting with the D7100, here are 100% crops 

D7100 f4 0.6s ISO 100

Between f4 and f8 diffraction is not a significant factor at all, so aperture size makes very little difference in sharpness.

D7100 f8 2.5s ISO 100

However, by f22 the width of the airy disk begins to overlap multiple pixels and multiply airy disks begin to interfere, severely limiting the resolution of the camera.

D7100 f22 20s ISO 100

The Nikon D50 is a much older, lower resolution camera, but as a result it has larger individual pixels spread out over the same sensor size. The result is that the camera has lower resolution at all apertures but loses less of that resolution to diffraction.

D50 f4 0.6s ISO 200

As with the D7000, there is little to no visible difference between f4 and f8 in the D50.

D50 f8 2.5s ISO 200

By f22 the 'fuzz' is quite obvious but the difference between f8 and f22 is much smaller than on the D7100 because the D50 has fewer, larger pixels.

D50 f22 20s ISO 200

Real world implications

Diffraction in APS-C and FX cameras is not as much of a restraint as it is in larger format cameras, but not understanding the concept will severely limit modern DSLRs that have very small pixels. You have to understand all the details or you won't get the most out of your camera.

Here's a stark example of that: Two images, taken with two different cameras through the same lens with the same settings on everything but aperture. The D50 is worth about $150, and the D7100 is worth about $1,100, but the D50 is set at f8 and the D7100 is set at f22. These images are raw from the camera, then I corrected the alignment, exposure (a tiny bit, they were pretty accurate already), and white balance. 

The camera on the left is the Nikon D50, the camera at the right is the D7100. This is a visual example of what many of the best photographers often repeat: the photographer matters more than the camera. If both were shot perfectly, the D7100 would win hands down, but if you don't understand the subtleties of photography you'll never improve just by buying new equipment.

Full resolution test photos 

Nikon D7100 f4
Nikon D7100 f8
Nikon D7100 f22

Nikon D7000 f4
Nikon D7000 f8
Nikon D7000 f22

Nikon D50 f4
Nikon D50 f8
Nikon D50 f22

DIY Macro: Reverse Lens Technique

Metadata: 18-55mm with reversed 55mm @ 55mm f36 1/60s ISO 250

High quality lenses with a 1:1 reproduction ratio suitable for true macro photography are extremely expensive. One of the best, the Nikon AF Micro-Nikkor 200mm f/4D IF-ED, costs about $1,794.95, so it's no surprise that the internet abounds with ways ways to cheaply modify lenses to increase the reproduction ratio. This page will show how to reverse and combine lenses to get really high reproduction ratios.

50mm f1.8, @ f8 1/250s ISO 100 with SB-800

First, however, I should mention that there are other, easier ways of focusing closer to your subject, such as these macro lens attachments.

This is the simplest way to increase your reproduction ratio. These lenses are usually very inexpensive and usually come in sets labeled +1, +2, and +3, like mine on the right. they simply screw onto the end of an existing

lens (I put them on my 50mm f1.8 prime) and reduce the focusing distance a bit.

D7100 50mm f1.4 @ f10 1/60s ISO 100 with three flashes

A way to get even closer, however, is to attach the lens backwards, either with a reversing ring (also known as a macro mount) or with electrical tape like here (though I was only shooting indoors with this). This lens arrangement allows you to focus much closer than with a set of macro lens attachments, but you lose the ability to use any lens automation—no camera controlled aperture, no autofocus, no distance metering (which only applies to D lenses)—so all that has to be set manually. As a result, this technique works best with lenses that have an aperture ring. Gelded lenses lack the aperture ring, so if they are detached from the camera they will only shoot at their minimum aperture, which makes focusing and framing difficult because the vewfinder is very dark. The reversed lens should always be focused to infinity, I found it helpful to tape the focus so it couldn't drift (unnecessary on lenses with internal focus motors as the internal motor should hold it in place)

Here's an example with a clementine (not a full sized orange): 

50mm f1.8 reversed @ f22 2s ISO 100

If this isn't close enough for you (and it wasn't close enough for me), you can actually combine multiple lenses. I attached my 18-55mm G lens to the end of my 50mm prime with a piece of electrical tape, making a lens that focused far closer to the subject and was actually durable enough to use without any worry of getting dust into the camera or having the lens fall off. 

D7000 50mm f1.4 @ f10 1/60s ISO 125

D7000 50mm f1.4 @ f10 1/60s ISO 125

D7000 50mm f1.4 @ f10 1/60s ISO 125

  This lens combination works really well for surprisingly sharp and close macros. The G lens is connected to the camera body like normal, so aperture control is retained. The inside lens should be zoomed to the furthest telephoto setting to reduce vignetting (and any filters attached to the lenses will increase vignetting) and focused to its closest focus point. The outer lens has to be a non-G stopped wide open (faster lenses like this f1.8 do best here) and focused to infinity. 

The biggest problem I had with this setup was the way that the lens can move in three different ways, outer focus, inner focus, and zoom, making focusing difficult. This problem was easily solved by simply taping across all three joints in the lenses so that it was always set to the closest focus distance. I focused by moving the camera and subject.

Here's a few examples!

18-55mm lens with 50mm f1.8 reversed @ f32 1/60s ISO 100 with SB-800

18-55mm lens with 50mm f1.8 reversed @ f32 1/60s ISO 100 with SB-800

18-55mm lens with 50mm f1.8 reversed @ f22 3s ISO 100

18-55mm lens with 50mm f1.8 reversed @ f22 1/125s ISO 1600 with SB-800