Nikon D800
Busting the Myth of Megapixels

Page 4

(June 2012)

     
 

Page 1 (Introduction, The test, Resolution, Chromatic aberration (CA))
Page 2 (Noise)

Page 3 (Picture Control, ADL and HDR, Vignetting control, Distortion control)
Page 5 (Nikon D800E, What do I find problematic with the D800?, Conclusion, Post Scriptum)

 
 
Diffraction and depth of field

As we all know, depth of field can be increased by stopping down the aperture, but at the same time we have to consider that from a certain aperture value on (commonly referred to as the diffraction limit) the picture starts to lose sharpness due to light diffraction on the edges of aperture blades. This phenomenon is present with all apertures, but when the aperture is open, it is so low, that no image sensor can detect it. Greater the image sensor's resolution (or the smaller his pixels), the more likely it will detect the undesired effect of diffraction. The D800's image sensor has very small pixels, so we can expect it will capture the diffraction effect with less stopped down apertures than cameras with larger pixels.

The diffraction limit can be approximately calculated by multiplying the pixel size (in mm) with 2000. Nikon D800's pixel’s size is 0.00488 mm (35.9 mm : 7360). When multiplying this with 2000, we get 9.75, which means that the aperture limit for the D800's image sensor is approximately f/10. In practice we can say that it's f/11 since we most likely will not spot a difference between f/8 and f/11. If we calculate the same for the D3 we find that its aperture limit is f/16.

I have to emphasize again, that it's the lens that is responsible for depth of field and diffraction, and NOT the image sensor. When using the same lens and the same settings (focal length and aperture) the depth of field and diffraction will be equal in all cameras, both effects will only be more underlined in cameras with higher resolution image sensors.

For the depth of field and diffraction test I chose a composition, where half of the frame showed a detailed structured book cover, which was focused by the lens and in the other half papers with text laying under the book. The book was 22 mm thick, therefore the text was 22mm behind the focus level. The photographed scene was 133 mm broad. I used the AFS Micro-Nikkor 60 mm f/2.8 lens. The photos were taken under equal settings (ISO 200, picture control: Standard, white balance: Preset) with both cameras (D800 and D3) and I used aperture stops from f/4 to f/32.
 
 

The upper crops with magnifications up to 100% fully confirm the theory. The book surface is equally sharp until aperture f/11 with the D800, at f/16 we see a slight unsharpness that becomes more apparent at f/22 and increases rapidly at f/32 so that the picture becomes practically useless. In photos taken with the D3 on the other hand, the book surface is equally sharp practically through the entire range of aperture stops. Only at f/22 we notice a slight unsharpness that slightly increases at f/32 however the result is definitely more useable than that of the D800. It seems as if the depth of field with the D800 is smaller than that of the D3. In reality it is of course completely equal, but the smaller pixels emphasize the unsharpness.

However this is all on a pixel level. Similarly as with noise and other aberrations that are more prominent in image sensors with a higher resolution (chromatic aberration, focusing errors, and camera shaking) also diffraction becomes less obtrusive when we reduce picture size. I reduced the photos taken with the D800 for 73% in Photoshop in order to make them of the same size as those taken with the D3. The diffraction effect reduced noticeably and became comparative with the one captured with the D3. In addition to that, the depth of field seemingly became equal to that of the D3 (it "increased").
 
 
In the end I photographed the test scene with the D800 from a greater distance, so that I captured a width of 230mm (to match the resolution ratio between D800 and D3). This way photos from both cameras captured objects with the same amount of pixels (similarly to the noise test). The depth of field seemingly became equal in both cameras (in reality it's greater with the D800 due to the greater distance, however smaller pixels make it appear smaller – the effects annul each other!), while the diffraction did not change and is thus more present in pictures taken with the D800 (since I did not reduce the size).
 
 
 

Autofocus

 

The Nikon D800 shares with the D4 the most modern Nikon autofocus system up to date. Basically a similar AF module with 51 AF sensors is used as in the D3 family, D700 and D300/S, but it had been improved in the field of light sensitivity. The old system operates up till the light value EV -1 while Nikon claims that the new system operates up till EV -2. The system is being controlled by a new processor that can process data faster and thus enable faster focusing. The old system is successful in focusing with lenses up to f/5.6, while the new system enables the use of lenses up to f/8, however only with the central AF sensor. For an underwater photographer this is a very welcomed feature when using teleconverters in combination with macro-lenses.

Most notably for the user, the user interface of the AF system has changed. The M-S-C switch next to the lens is gone, so is the button to choose different AF sensors patterns in the camera's back. Instead there is a single switch (M - AF) next to the lens that includes a small button in the middle. This button in combination with the main command dial and the sub-command dial enables us to change several camera settings. With the sub-command dial you can switch between S (single) and C (continuous) mode, while the main command dial enables a quick choice of AF sensor patterns that are now more numerous than ever. For an old Nikon photographer this might be inconvenient at first, considering that we have been using the M-S-C system for more than two decades, back in the good old film times. But we get acquaint quickly to the new system and soon we discover its advantages. We have the entire control over the AF system without the need to move the camera away from the eyepiece, seeing as all the information is available through the viewfinder.
 
 

I did not test the autofocus comparatively with other cameras. To tell the truth, I'm still getting familiar with it and I'm discovering the differences between modes like the "51-point dynamic-area AF" or "3D-tracking" and "Auto-area AF", which all use the entire 51 AF sensors.

In order to test the AF system I asked my model to run towards me as fast as possible. The model was intentionally dressed in red, as the camera considers colors while focusing. I used the AF-S 70-200 mm f/2.8 II lens at 200 mm, aperture f/5.6, exposure set to A, image size 1.2X, shooting speed "High" (5 frames per second) and ISO 800. Obviously I chose C (continuous) as the focusing mode and tried different AF sensor patterns.

At the beginning I was disappointed with the "Auto-area AF" mode, discovering that more than a half of the shots taken were unsharp! Then I realized that I have to show the camera what to focus by pointing one AF sensor on the object. Then the situation changed substantially better and the AF sensors did a splendid job following the model while moving. From 24 shots that I've taken only 5 were unsharp (sharp photos are framed in green, unsharp in red).
 
 

The 3D-tracking modes consider the object's color while focusing. The camera immediately locked the focus on the red sweater and followed it with the AF sensors. Out of 24 shots only 4 were unsharp.

 
 

I also decided to test the 9-point dynamic-area AF mode, in which I pointed all 9 AF sensors manually to the face of the model. This was the least successful mode since 6 shots out of 19 fired were unsharp.

 
 

None of the photos ware entirely out of focus and all unsharp photos had a focus which was slightly in the background (i.e. on the ear or even further back instead of on the eyes). However, a successful focus depends greatly on the photographer's skills and here my lack of experience comes to show. Generally speaking I was under the impression, that the new AF system truly is more advance than the old (in the D3).

I was most fascinated by the 3D-tracking mode, considering that the AF sensors literally glued themselves on the red sweater. The new system must be highly useful in sports photography, but sadly less so in nature photography, because most animals try to fit into their living environment with both pattern and color.
 
 

White balance (WB) and monitor

 
Like most digital SLR cameras, the Nikon D800 enables several options for white balance (WB). I personally am only interested in the following modes: A (Auto), Pre (Preset) and K (color temperature in Kelvin). Experience shows, that cameras usually have the most difficulty in the automatic WB setting when shooting under fluorescent lights. In such cases the Preset mode is by far the best solution. For this test I used all my three current cameras: D2X, D3 and D800. I closed the blinds and prevented any daylight to enter the room, and then I turned on fluorescent lights and took photos of an entirely white wall. All three cameras did a great job in the Preset mode and reproduced the whiteness of the wall with exactly the same result (even the 8 years old D2X!).
 
 
In the Auto WB mode however, things got a little bit trickier (as expected): all three cameras failed the wall test, the D2X gave a much worse result than the other two cameras. The D3 and D800 were substantially closer to reality than the D2X, yet the result still had a strong yellow hue. On the camera monitor the D3 showed a reddish hue and the D800 a yellow-greenish hue.
 
 
 

But when I compared the photos on the computer monitor I was very surprised to discover that the photos from the D3 and D800 are much more similar than it seemed on their camera monitors! This means one of the monitors must be lying, but which one? I've noticed complaints about the monitors of the D800/D4 being too yellowish on the internet and I've read Nikon's official response that says its colors are truer than the colors in monitors of the D3 family... What is true? At first glance the solution to this seems simple and obvious – I should take photos of both camera monitors and compare them with the photo from the respective camera on the computer monitor. But what WB setting should I use to get an objective picture of the monitors?

I decided to proceed as follows: On the D3 test camera I first made a 100% white "photo" (I shot a 4 seconds long exposure at daylight with no lens) and played it on the camera monitor. Then I used the Nikon D2X with a AFS Micro-Nikkor 105 mm f/2.8 lens to measure the WB in Preset mode and calibrated it this way to the color temperature of the D3 monitor. I played the photo of a white wall taken in Auto WB mode on both cameras (D3 and D800) and took a photo of their monitors. I repeated the procedure with the D800. This way I shot four photos with the D2X, two of each camera's monitor with two different WB calibrations. I slightly changed the brightness of these four photos in Photoshop, but I did not edit the colors.

I reduced the size of the photos that portrayed the white wall from both cameras to 420 x 280 pixels (two copies per camera) and pasted the photo of the corresponding camera's monitor in the middle of this photo (size 270 x 180 pixels) in order to make comparison easier and more objective. On the wall-photo taken with the D800 I've pasted two photos of D800's monitor each shot with a different WB calibration, and on the wall-photo shot with the D3 I've pasted two photos of D3's monitor also shot with two different WB calibrations.
 
 
 

One would expect that the photo made with the camera would be most similar to the photo of the monitor, when shot at the WB setting calibrated to this same monitor. The photos above confirm this assumption, but at the same time we can see that the photo taken with the D800 is slightly closer to the photo of its monitor (upper left picture) than this is the case with the D3 (lower left picture). This means that Nikon is right. The monitor of D800 indeed has a slight yellow-greenish hue, while the monitor of the D3 displays a slight orange-reddish hue; however, the color deviation is smaller in the D800.

 
 
Page 1 (Introduction, The test, Resolution, Chromatic aberration (CA))
Page 2 (Noise)

Page 3 (Picture Control, ADL and HDR, Vignetting control, Distortion control)
Page 5 (Nikon D800E, What do I find problematic with the D800?, Conclusion, Post Scriptum)