This is the 54th in a series of posts on the Fujifilm GFX-50S. The series starts here.
When I was testing the Zeiss 250/5.6 Superachromat on the GFX yesterday, I noticed that the combination of the camera base mounting and the long focal length made it quite sensitive to vibration. I thought that it might be a good lens to use to test the GFX for shutter shock.
Shutter shock on the GFX? Doesn’t it have EFCS? Yes it does, but EFCS won’t operate at shutter speeds faster than 1/500 second. Above that speed, the camera switched to the mechanical shutter, unless you’ve selected the fully-electronic shutter.
Let’s review where shutter shock comes from, and why it can be especially pernicious with mirrorless cameras. Shutter shock is a name for vibrations caused by the camera’s shutter that affect the sharpness of images. It is related to an effect that occurs in single lens reflex cameras called mirror shock. Mirror shock is the result of the motion of the mirror in an SLR swinging up and out of the way so that the exposure can commence. It has long been recognized as potentially problematical, and most SLR’s have ways to trigger the mirror independently of the shutter so that it doesn’t ruin photographs where such vibrations would damage images.
In a mirrorless camera, there is no swinging mirror to induce vibrations. However, with a mechanical focal plane shutter, there is another problem. To gather the data necessary for the LCD display on the back of the camera or the electronic viewfinder (EVF), the shutter needs to be open. When the user presses the shutter button, the shutter winds, which closes it, and the first curtain is immediately triggered, opening it again. In cameras like the Leica M240 and the Sony a7R and a7RII, the winding of the shutter is what causes most of the problem. The Leica M240 allows v=liveview to be turned off before the shutter is tripped, avoiding the problem much like “mirror up” in SLRs. The a7R has no defence against shutter shock, except avoiding the shutter speed where it is most likely to be a problem. The a7RII has a way around shutter shock: electronic first curtain shutter (EFCS).
In mirrorless cameras with EFCS, there is no need to wind the shutter before the shot. The mechanical first curtain is replaced by an electronic rolling reset signal that causes no vibration. The only vibration is that caused by the second curtain, which ends the exposure. That is usually not a problem. However, EFCS doesn’t work well at high shutter speeds, where the timing of the mechanical and electronic curtains may not be the same, and where the fact that the mechanical curtain is in a different plane than the electronic one can cause odd things to happen. Nikon does not allow the use of EFCS at shutter speed faster than 1/2000 second. In the a7x cameras that support EFCS, it is possible to use it at high shutter speeds, but Sony warns that unacceptable results may occur.
In the GFX, the mechanical shutter moves more slowly than those in the a7x, and, because the cover glass is well away from the sensor, the mechanical shutter’s plane of operation is farther away from the sensor than in the a7x. These differences conspire to lower the fastest speed where EFCS is safe to use. Fuji takes Nikon’s dictatorial approach and will not let EFCS operate at shutter speeds faster than 1/500 second.
In my testing of the a7R, I noticed shutter shock at 1/640 seconds and slightly higher, so it seemed that it would be prudent to see if the GFX had vibration problems at that speed.
I put the GFX in landscape orientation on an Arca Swiss C1 head attached to a pair of RRS heavy-duty carbon fiber legs. I aimed the camera at my backlit razor blade and turned the Wescott LED panels up as far as they would go. Color temperature was set to 5500 K. I got as close as I could with the Superachromat, which is about 10 feet. With the lens wide open, I focused as best I could on the razor blade, which was pretty sloppy, both because it’s hard for humans to focus on that edge, and the image was jiggling around every time I twisted the focusing ring. I figured it wouldn’t make much difference, since I was just going to look for changes. I set the camera’s shutter to 1/500 second, and the shutter mode to EFCS. I turned on the 2 second self timer. I made 8 exposures. Then I turned the shutter speed to 1/640 second, which turned off EFCS and gave me fullly mechanical shutter. Eight more exposures. I developed the images using dcraw with the Imatest defaults, and measured the MTF50 in cycles per picture height. Here’s what it looked like:
The middle of the columns, at the top of the orange lines is the mean, or average value of MTF50. The lower end of the orange is the mean minus twice the standard deviation, and the top of the gray is the mean plus twice the standard deviation If the statistics turned out to be Gaussian, 96% of the outcomes would lie between those extremes. Clearly, turning off EFCS hurts sharpness, even at 1/640 second. Also clearly, the effect is immaterial to normal photography.
Next I fitted the Fujifilm 120 mm f/4 macro lens. I made the same set of exposures, with two differences: I let the camera’s autofocus system do the focusing, and I made 16 exposures at each shutter speed.
The results are sharper. That’s because the Fuji 120 is a sharper lens than the Superachromat, and also because at f/4, the Fuji GFX can focus on the razor blade better than I can. The standard deviation is also larger. That’s because vibration makes more and more difference as the lenses and sensors get sharper and sharper. EFCS is better. The mechanical shutter is worse. The difference is unimportant.