I have confirmed a loss of sharpness using the Sony FE 55mm f/1.8 Sonnar under ideal (for the camera) studio conditions at ƒ/4, varying lighting intensity, but keeping the aperture constant and the focus on manual (of course). The lens is superb.
Sony A7R Shutter vibration is NOT just a telephoto lens issue, or an issue with adapted lenses.
I have done all the back and forth and cross checks, all the rigor (including a cross-check few would think of). Proofs will follow; I am preparing a detailed analysis to present in my review of the Sony A7R in Guide to Mirrorless.
Field conditions cannot be better than this studio setup, only worse. And Murphy’s Law says that the shot you worked for all day will be affected. Certainly has affected some of my scenes, which makes it awkward to comment on the lens (is it the lens?!)—makes me have to be extra cautious, but peak sharpness can never be discounted, only oddly unsharp results are suspect of course.
And of course many frames will have no problem at all at higher shutter speeds, nor do I know if handholding somehow dampens the vibration (enough or at all). Some readers seem to think handholdng avoids the issue, a theory consistent with my mass-coupling idea in How to Hold a Camera Steady how-to article in MSI.
The 24-megapixel Sony A7 has none of these issues: bulletproof consistency due to its EFC option (electronic first curtain).
Firmware update to mitigate?
Sony PR informs me today that Tokyo has been informed of the issue, that Sony is currently investigating this issue, and will be in touch with me for a discussion. So I have my fingers crossed that some mitigation is possible.
This is how the A7R currently operates the 2-second self timer:
- Release shutter (via wireless remote to avoid disturbance from pressing).
- Live View remains active for duration of timer, e.g. 2 seconds.
- Camera slams shutter closed, then immediately opens shutter for the exposure.
This is operationally flawed.
I am hopeful that the mitigation strategy described here, if not entirely effective, will be at least partially effective. The A7R should immediately close the shutter. Then the vibration from the shutter closing would dissipate in the ensuing 2 or 10 second delay (Sony ought to offer configurable delay BTW). With this delay, the shock wave has subsided; the shutter then opens to make the image (which can cause its own vibration, as can the closing curtain). But at least the first shutter-bang is long gone (choice of 3/4/5 second delay would be very helpful for longer lenses, 2 sec too short, 10 sec too long).
In self-timer mode, or a to-be-added shutter delay mode, a proposed mitigation:
- Release shutter.
- Camera immediately closes the shutter; vibration dies down for the duration.
- After self timer or “anti shock” delay, shutter opens to make the exposure.
It’s unclear if this would eliminate the problem, but it would almost certainly reduce it. And it does not address handheld shooting, since the Live View would black out.
Olympus offers an anti shake mode on the E-M1 and E-M5 cameras which my understanding is does exactly the above. Use of Classic mode in the Leica M240 offers this behavior also, and it helps (does not solve, but helps).
Joseph Holmes on adding mass to the A7R:
I agree that the self timer mode (with options for at least 2, 3, 4, 5 and 10 seconds, if not all whole-second times from 1 through 12 or even 20, must have the close-the-first-curtain-early option available to them. At least so that tripod shots with self timer can be sharper.
But it's also vital that we have a mode which is simply like MUP mode. Press once to close the shutter and press again to take the picture. This only works when using a remote (avoiding finger motion on the button), but it also allows us to choose, at the last moment, when to trip the shutter.
Precise timing matters very often with anything that's moving, and a Shutter Delay mode would just be fantastic for remote use situations (whether IR or wired or some future radio operated one like the Vello ones which don't currently include one with a cord of the right type for the A7 family).
The photographer would need to know how long to wait for the second press, but I'll bet you that a very short delay will do the trick, and they will learn how long to wait, just like they did with MUP mode. How else could it be true that at 1/100th, the effect is far worse than at 1/10th? That alone proves that the dampening is mostly very quick (given that the second curtain's almost certainly not doing it). A delay of 1/4 second should be plenty. Click - click. Done.
Biggest drawback is double the number of presses of the remotes, wearing them out and using up the IR remote's battery twice as fast. No big deal.
On a related topic, Michael Schultz and I are close to having worked out and built a damping weight which will easily attach to the camera with a clamp. Start with the RRS L-plate (due to start shipping the earliest orders within two to four weeks). Make a damping weight which consists of just three parts: The RRS flat-backed clamp B2-FAB-F; a SS screw from RRS, the 1-inch, 1/4-20 Cap head screw, and lastly a block of either Brass or Bronze of Stainless Steel (there are many alloys of each).
I chose 316 Stainless, 1.25" front to back, 1" tall, and cut to 3" wide from bar stock. The block weighs 17 oz. The clamp weighs 4, and the L-plate weighs 3.
The block will have a shallow groove milled in it, at least .060" deep, to prevent the clamp from twisting. It will also have a hole drilled, 3/16" off center to the rear, to position the center of the block's mass beneath the shutter, 3/16" forward of the tripod screw hole, and counter-bored to enlarge the lower end of the hole to fit the allen head of the screw properly, have the surfaces cleaned up with the mill and the edges chamfered and eased.
The total weight will be roughly 23 to 25 oz including the clamp and the L-plate, depending on the alloy chosen. OnlineMetals.com can supply the piece of Stainless, cut to the approximate length, for about $27 shipped to central California. I chose Stainless over brass because I prefer a neutral color for my camera gear. Brass and bronze are slightly more dense.
The clamp will not fit if some adapters are used, e.g. the Sony LAEA 3 or 4 and maybe not the Metabones adapter. The front edge of the clamp will bump into the foot of the Sony adapter by less than 2 mm. The RJCamera adapter has an optional collared foot which can rotate, like the Novoflex collared foot, so a vertical camera orientation would mean that the foot would be down and the deadweight would then be to the right (i.e. it would work).
The dead weight would also work fine with lens tripod feet and when a camera is in vertical position with a native lens. Or in vertical position with your Novoflex collar and foot. Or with any Mirex foot in either vertical or horizontal position.
For horizontal images on a Metabones or Sony adapter foot, a weight can be made for direct connection to the camera (no L-plate and no clamp, just screwed into the camera), but this would only work OK if no native mount lenses were in the outfit which required direct mounting to the tripod via a camera plate (the block would be in the way of the plate).
I did initial experiments with 26 and 13 ounces of weight for the horizontal camera position. The heavier weight was highly effective, the half-weight was quite insufficient. Mike's experiments have shown a very high degree of effectiveness of a 24 ounce weight screwed directly to the bottom of the camera for both horizontal and vertical frames at 180 and 280 mm with a 70-200 Nikkor on a Vello foot without, then with the Nikkor 1.4X tele-extender, respectively. At 1/100th (the typically most problematic shutter speed). It works.
Let's hope Sony's fix does more than half remediation. They will do this. It's bad, they will know it, they can easily do the firmware fix. I have a lobbying campaign underway too. I like your testing protocols. They're very good.
DIGLLOYD: SCM (shutter close mode) sounds good to me too.
I agree that the shock wave appears to be a very high frequency and very quickly damped. It suggests that one one workaround on a tripod is to “go slower”, e.g., use a neutral density filter if necessary, or a polarizer. But both approaches change the image, with ND filters increasingly yellow with too much density.
If the Sony A7R gets fixed, it will represent the best digital on the market today at the best price with the widest array of lenses. It won’t be appropriate for all shooting situations, but only a little imagination is needed to see where this is going, and 54 megapixels or so is surely a year or two away at most. Some attention to shutter damping is in order for future models too.