No, this is not specifically about moving images. So shoot me. It’s really sweet, and I had to post it.

Wired is reporting on a experimental plenoptic still camera outfitted with 90,000 microlenses that enable changing the focus point of a captured image in post production, even if nothing in the original captured image was actually in focus.

In an overly-simplified-nutshell, it works by capturing not just the light that exposes the CCD, but the angle and direction of the light rays. And that enables the image to be reconstructed in post production with a different point of focus, or even allows shifting the POV of the viewer across the aperture of the camera lens (effectively allowing you to “look behind” something in the image). It’s pretty incredible stuff, and you have to see it to believe it.

There is a detailed writeup at the Stanford site, gallery of images, and even a video that explains exactly what is going on. Check it out.

Quick notes:
The camera can operate on very short exposures. As an example of this, the video and PDF include a shot of a drop of water, caught in mid-splash. The camera does not capture images that are as sharp as a conventional camera at the same exposure setting, however, the signal-to-noise ratio of the light field images is much higher than the conventional images. 5.8x the SNR, to be exact, due to the light field camera capturing 25-36 times the light for a given image.

I’d be curious to hear what ways this can/will be applied to moving images. What kind of storage requirements are we talking about, and how quickly can an image be taken. Is 24fps possible? Imagine a camera that would allow you to capture not just the exposure and color characteristics in LUTs (Look Up Tables), but also allow you to manipulate focus in post production…