Sharpness and the D850

Recently I switched from a 12 MP Nikon D700 to the 46 MP Nikon D850. Along with the 4 times increase in resolution come new challenges. An obvious one is lens sharpness – you’ll be surprised how many “sharp” lenses aren’t that sharp on the D850.

Yet, soft lenses are the least of the problems. If you want to have reasonably sharp photos, you will have to hone your technical photography skills. For any kind of landscape work, the tripod will become your best friend. But often enough you may just want to grab your camera and take photos.

Recently we went to Europe on vacation. I took the D850 paired with the Tamron 15-30/f2.8 and the Nikon 70-200/f4, and the Novo Explora T5 carbon fiber tripod. The backpack was my new Peak Design Everyday Backpack 20L.

First we visited Berlin, Germany. During the day we walked around the city. I felt that carrying a tripod wasn’t really needed as the light was good. Fast forward – I’m back home now and going through the photos I discovered the following oddity.

D850 photo sharpness
French Cathedral (Hugenottenmuseum) at the Gendarmenmarkt, Berlin (Nikon D850 with Tamron 15-30/f2.8 @20mm f8)

The above photo was taken at 20mm focal length, 1/320 second, aperture f8, at base ISO 64. Vibration control was turned ON. (The photo has been processed in Lightroom for white balance, exposure – the camera underexposed by 1 stop – shadows and highlights, but no perspective corrections.) At this size you won’t be able to notice anything wrong – right?

Below is a detail from the bottom right of the image, but far away from the corner or edges (see the original photo above to locate this detail).

Nikon D850 sharpness
Detail of the French Cathedral photo – the left side is sharp, the right side blurred

The foreground (cobblestone pavement), concrete bench and the bushes, as well as the building on the left are sharp. But when you move your eyes to the right everything gets blurred!

Here is another detail from the left lower side of the photo.

Nikon D850 sharpness
Detail of the lower center left of the French Cathedral photo, with the left side blurred

Notice how the picture above looses sharpness from right to left (see the blurred flower pot). Here is another example from the left side of the French Cathedral photo.

Picture sharpness
Detail from the left side of the French Cathedral photo – notice the sharp urn (acroterion) on the right, and the soft one on the left

The pediment of the building above is adorned by two urns or acroteria. The right one in the photo is sharp, whereas the left one is blurred. The tiles on the roof behind are all sharp, though.

All those soft areas in the photo are starting to appear about 1/4 width from the center.  This means the blurred regions start too far away from the edges to be simply attributed to “corner softness” of the lens.

The depth of field at 20mm focal length and f8 aperture, when focused at a distance of about 4m, is from 1.3m to infinity. The pavement in the bottom center of the frame is quite sharp, and about 1/8 away from the edge towards the center it looks perfectly fine. The following is a Google map satellite image of the place I took the picture, with the lines showing the approximate location and direction the photo was taken.

A Google Maps satellite image of the approximate location where I took the above picture

Looking at the satellite image above you can easily see that the blurred areas are not out of focus. Notice the red roof on the left with the two “urns” and the top of the roof? If it was a matter of subject distance (i.e. out of focus), the third decoration on the top of the roof would have to be as blurred as the left urn, as it’s about the same distance from the camera. But it looks sharp.

So if this blur is neither corner softness nor out of focus, what is it? What’s more perplexing, the center of the photo looks fine.

Like every lens, the Tamron 15-30/f2.8 G2 does have a little softness in the corners or perhaps edges, but nothing close to what I see here.

I like to narrow the possibilities to the following two:

  • Vibration reduction – a fault of the lens (or mine, since I haven’t turned off VR)
  • Circular movement around the lens axis – sloppy technique on my behalf

Motion blur at 20mm focal length and a shutter speed of 1/320 second seems an unlikely cause. At 1/320 second you could freeze the action of many sporting events, even with a medium telephoto lens. Yet I tend to believe that it is exactly that – motion blur!

Looking at the photo, it seems that the blur increases with the distance to the center of the image. It’s like a circle around a relatively sharp center image. In other words, I have been turning the camera around the axis of the lens when pressing the shutter release.

In this case, the center of the image moves relatively little compared to the outer parts. This is because the distance a given pixel moves depends on the distance to the axis around which the movement takes place.

Here some calculations. Let’s assume I had twisted the camera around the center of the sensor at a speed of 5mm/sec for the pixel that lies at the center of the left (or right) edge. At a shutter speed of 1/320 sec we get the following pixel movement:

5mm=5000µm

5000µm/sec / 1/320 sec = 15.625µm

The Nikon D850 has a pixel pitch (distance between pixels) of 4.34µm. In the example above the outermost pixel would move a distance of

15.625µm / 4.34µm = 3.6 pixels

The above calculations are approximations (I calculated linear movement, not circular movement). Let’s see how this translates into angular velocity.

Nikon D850 sensor size: 35.9 x 23.9 mm

Circumference (for the outermost pixel at center height):

35.9mm x π = 35.9mm x 3.14 = 112.7mm

A movement of 5mm translates to an angle of:

360 degree / (112.7mm / 5mm) = 15.9 degree

In other words, with a high resolution camera a small twist of 16 degrees per second translates into a substantial pixel shift at the outer areas of the image.

The above figures are hypothetical. I wouldn’t know if I twisted the camera at 15.9 degrees per second or more or less. But 16 degrees per second is neither slow nor fast, and definitely possible.

Now here is what happened, as I recall the situation:

I had taken the camera down and looked around to make sure that nobody would cross the square in front of me. When the square was clear I lifted the camera to my eye and twisted the camera to get a straight horizon. I was in a hurry to take the picture as the boy on the electric scooter on the right side of the frame was approaching fast. The closer he got the more prominent he would be in the frame, and that wasn’t what I wanted.

I must have pressed the shutter release while still framing the horizon. In other words, I twisted the camera when pressing the shutter release.

Of course my family starting to loose patience didn’t help. If it had been important to me to get a good shot of that scene, I would have spent more time in planning and execution.

Afterthought (January 24, 2020)

I continued using the Tamron 15-30mm and the above phenomena reappeared in more photos. Sometimes it is only one side that is blurred.

It is also not predictable. One shot may be fine, the other lacking sharpness, usually on the right side. I will have to send in the lens.

Repair 1

In March 2020 I eventually turned in the Tamron 15-30mm for repair. Due to CoVID lock-down, it took until May to get it back. As a replacement I purchased a Nikon 24mm/1.8. It doesn’t have VR, but I tell you, this is a nice and sharp lens. Also smaller and lighter to carry. But it can’t replace a true wide angle lens or zoom.

The repair slip for the Tamron read that they had done some adjustments/calibration. When I took it on a hiking trip it turned out it was worse than before. So I returned it for repair.

Repair 2

I was able to pick up the Tamron 15-30 G2 at the end of October, 2 days before another hiking/climbing trip. This time we visited a cave, the longest salt cave in the world to be precise. You can guess that the conditions for taking pictures were difficult. Below are two outdoor shots as we climbed Mt. Sodom to the cave entrance. That was the easiest part.

Upper Malham cave
« 1 of 2 »

I had only one camera flash with me and the D850 lacks an internal flash. So no remote trigger, only on-camera flash which isn’t good. (Unfortunately I misplaced my remote wireless commander.) The photos below are heavily edited for exposure correction, shadows and some noise reduction. This is hardly a valid “test” for a lens, but they may give some indications as to the performance.

Upper Malham cave
« 1 of 6 »

So far the photos look good. Our next hike is in early December and I plan to give the Tamron 15-30/2.8 G2 another chance.

Conclusion

Here is what I take from this and other experiences:

  1. Test your lenses before putting them to work
  2. The D850 is no snapshot camera – take another camera if that is what you are going to do
  3. High shutter speed doesn’t necessarily prevent motion blur
  4. Steady the camera BEFORE pressing the shutter release
  5. If a shot is important to you, take the time for planning and execution
  6. Use a tripod whenever possible
  7. Turn off vibration reduction at higher shutter speeds or when using a tripod
  8. Practice, practice, practice…

Please leave a comment and share your opinion/experience.

Heiko Sieger

Author: Heiko Sieger

The day has 24 hours. If that isn't enough, I also use the night.

This site uses Akismet to reduce spam. Learn how your comment data is processed.