The purpose of this test was to measure the time and amplitude of shutter vibration on a Nikon D200. Discussions about mirror lock up, Exposure Delay Mode and other approaches to eliminating vibration abound on the internet. Does mirror lock up actually matter? I wanted to see what the real story was.
This test was done with a D200, an 85mm f1.4 with screw on lens hood placed on a BH-3 from Kirk photo, supported on a Bogen 3 section aluminum tripod on a carpet. The camera was set to an exposure time of 1/4 second. The shutter was fired with a cabled 10-pin electronic shutter release, with the wire tied down to an independent light stand, eliminating detectable vibrations coupled in from me depressing the remote release connected to the cable connected to the camera body. An Analog Devices ADXL210 accelerometer, soldered to an FR-4 PCB, driven by a 4.91V supply, was mounted to a Nikon AS-15 hot shoe synch cord mount which was attached to the D200’s hot shoe mount.
This test was shot at 1/4 second. Fortunately the oscilloscope measurement indicates that my shutter is firing accurately in the time range – an added bonus.
Figure 2: Mirror slap up and shutter opening.
The mirror up time from when the shutter is depressed to actual exposure beginning is 50ms, the first vibration waveform. The second vibration waveform is the shutter opening click which lasts less than 30ms. Incidentally, this data confirms the stated shutter lag time of the D200 of 50ms between the mirror going up when the shutter release is depressed and the shutter actually opening, beginning the exposure.
Figure 3: Shutter opening vibration during actual exposure.
My particular D200 exposure delay mode: 500ms (data sheet shows 400ms from mirror up to shutter open)
Mirror up vibration time: 50-80ms
Shutter open vibration time: 28-30ms
Mirror down vibration time: 150-160ms
Not shown in the above diagrams are the tests done with Exposure Delay Mode, where the delay time of 500ms was determined. Also, a test was done with mirror lock-up and the exact same vibration pattern was observed as Exposure Delay Mode, only the shutter click vibration occurred much later, when I depressed the shutter release the second time. It was difficult to capture MLU (mirror lock up) and Exposure Delay Mode on the oscilloscope with enough resolution since the times were so long. However, the same vibration events were illustrated by an exposure time of 1/4 second. In fact, the same spike pattern was observed in all three types of exposures – a 1/4 second exposure, Exposure Delay Mode and mirror lockup. This gave me confidence that there was nothing else unaccounted for inside the D200 during these different exposure modes.
As can be seen from figure 2, the mirror up and the mechanical shutter opening are two independent events. These events are separated by 50ms. However, when firing the D200 in normal shooting mode (with Exposure Delay Mode disabled), the two different sounds (e.g. vibrations) are indistinguishable. However, with Exposure Delay Mode activated, there are two distinct sounds separated by 500 ms, as measured on the oscilloscope. The first sound is comparatively loud and the second sound is a quiet “click”.
Figure 3 is interesting because it shows the the shutter opening vibration, independent of the mirror up vibration, with this camera body, lens, ball head, tripod and surface configuration. This vibration lasts from 28-30ms, in the 1/30 to 1/40 exposure realm. This means that during an exposure of 1/35 of a second, the camera body is vibrating the entire time. Not much, but the vibration is there never the less. However, this does not mean photographs made in the 1/4 to 1/30 second range are always going to be blurry. That is a logical leap. I have made tons of sharp photographs in this shutter speed range. But from discussions I have read, this has always been the most difficult hand-held shutter speed range to achieve sharp pictures.
Why the 1/4 second value? I had to think about that for a moment. 1/4 second is about 8 times longer than the measured vibration time. Exposures longer than 1/4 second relegate the vibration to less than 10% of the exposure time, making the vibration insignificant to the image. Another way to look at it 1/30 of a second is three stops less exposure time than 1/4 of a second. The time the image is exposed while the camera is vibrating contributes very little to the image.
Don’t believe it? You don’t even have to do this with 1/4 second to 1/30 second exposures. You just want to see the effect of three stops less light on your image. Put your camera in manual mode, find the correct exposure of your image and then shoot it. Now reduce the exposure time by 3 stops (8 times faster) and shoot the image. That very underexposed image is what’s effectively recorded when the camera is vibrating – not too much of a contribution. If you could somehow take away the vibration time from your exposure time, you would go from 1/4 second to 1/5 second and only have your image underexposed by 1/3 of a stop – almost nothing.
One interesting thing to note is when I hand hold the camera in low shutter speed situations, the Exposure Delay Mode (of 500ms rather than the states 400ms) vastly improves my keeper ratio. The blur comes from my initial depression of the shutter release button. The vibrations I induce when pressing the shutter exceed the 50ms lag time between depressing the shutter button and the actual exposure taking place, adding blur to my images. Using the Exposure Delay Mode when my exposures drop below the 1/focal length time has worked out quite well, allowing the vibration from my hands to cease before the shutter is opened.
Interestingly, pressing with a large amount of downward force on the tripod or the camera with my hands did not appear to affect the vibrations at all. My guess was that the vibrations would dampen out a bit but they did not, neither in time and amplitude. This is not what I expected. My hypothesis was the vibration will dampen out faster and be of lower amplitude with my holding down the camera body with significant force. The hypothesis was proven wrong in the data. This shows the internal acoustic vibrations transferred through the body up to the hot shoe mount can’t be stopped by me hanging on. Perhaps mounting the D200 on a large block of granite will dampen these vibrations more but who wants to carry around a 150 pound rock to make images with? And the tools it required to correctly mount the camera to that huge rock are for an academic lab experiment.
When you have engineering training and access to the right equipment, you can run all sorts of interesting experiments. I always wondered why people say the 1/4 to 1/30 second range is the most difficult for hand-holding, ignoring advanced is Nikon VR (and Canon IS) technology. Now I have my answer. But, what about exposures faster than 1/60 of a second? I don’t have an adequate exposure explanation as above, so I will wave my hands and suggest that the exposure time is fast enough that the vibrations are mitigated. A logical leap? Sure. The best I can do is suggest that the large scale vibrations induced by a steady, human while hand-holding the camera are huge compared to the vibrations induced by the shutter opening. The large scale swaying is usually made insignificant beyond the traditional 1/(lens focal length) shutter speed rule. This argument is not totally consistent with the above data but it will do for now.
This data also suggests that with smaller lenses and decent tripods and heads, mirror lock makes no difference in the sharpness of an image. When the shutter is opened independent of mirror slap up, there are vibrations induced into the camera. Mirror lock-up does not help this. On larger lenses, less sturdy tripods or in other situations, mirror lockup may be handy. But with the D200’s Exposure Delay Mode ability, why would one want to mess with mirror lockup mode? If the vibrations from mirror slap-up do not settle down in 500ms (0.5 seconds), you probably need a better tripod and head.
The only advantages I can see is it’s faster to use a dial to select mirror lockup mode rather than having to go through a menu to find Exposure Delay Mode. That’s fairly minor since you will be taking more time to shoot this anyway. And you have to using an electronic cable release to make this worthwhile anyway. And once the mirror is locked up, you can’t see where you’re shooting anyway. The other advantage would be to cut down the 50ms lag time to basically no time. Does that really matter since the fastest humans have a reaction time of 100ms? And you can’t see where you’re shooting. But, it’s nice to have the ability anyway. I like the Exposure Delay Mode because in 0.5 seconds, I have my vibrations damped out and I don’t have to hit the shutter release again.