( slide your mouse over the image side to side for animation ) Still Loading: [ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ]

(The white line on the left side of the blue “exposure” band, below the big “@”, is when the camera's flash fired for the exposure.)

Marianne Oelund recently posted a sequence of images of a Nikon D3 shutter in action, over on the Digital Photography Review Nikon D3 forum. With her permission, I have turned that sequence into a mini web movie application, such that you can animate the frames by sliding your mouse back and forth across the image.

I thought she had created the images with a high-speed camera, but it turns out that, lacking a high-speed camera, she went to great (and ingenious) lengths to create this sequence. In her own words:

“ The setup used to create the images was rather simple; anyone with basic macro equipment could do the same. It just requires quite a bit of patience! I took a total of about 200 frames, and selected 70 for the slides. The imaging camera was another D3 (but any camera could be used), with a Micro 60mm lens, a white reflective collar around the lens, and a downward-firing SB-900 mounted above. Cable releases were used on both cameras, to avoid disturbing their positions. What wasn't quite as straightforward, was collecting the timing data. This is the area where a little engineering knowledge and equipment comes in. The “subject” D3 camera had an SB-800 mounted, set to manual 1/128th power, and turned toward the rear. A phototransistor detected this flash pulse, which was captured on an oscilloscope. Another phototransistor was used to sense the imaging camera's flash pulse, then the 'scope could give the precise time delay between the two. From the timing data, one can calculate the mechanical velocities, and even accelerations. It's worth doing this, as the numbers are impressive, to say the least!

I love this kind of ingenuity! A high-speed camera would have been so much easier, but hey, if you don't have one, but do have an oscilloscope and a couple of phototransistors (and the knowledge and patience to use them), Marianne shows you can get great results.

I took Marianne's 70 images, converted them to grayscale, cropped and shrunk them, added the various annotations you see on them, and wrapped them all up in a mini web application similar to the Kyoto cherry-blossom timelapse that I posted in the spring.

In looking at the actual shutter (between the mirror and the sensor), it seems that it takes about 3 milliseconds to drop the 26 or so millimeters it needs to expose the sensor. That works out to an average speed of about 8.7 meters/second, or just over 19mph. That's a lot of speed to accelerate to and from so quickly. Wow. Marianne has more timing analysis in her dpreview thread.

If you'd like to know more about how a shutter works (and, in particular, why the shutter design and operation has such an impact on how flash photography can be done), see Derek Miller's excellent writeup on the topic, linked from my “Flash Sync Speed and Other Mysteries“ post.

(The individual frames for my version of the movie above are hosted at Flickr to reduce load on my server.)