Rob’s recent post over at Boing Boing drew my attention to computer scientist Ren Ng’s startup, Lytro, which is developing a digital camera technology that requires no moving parts to focus an image at any depth of field, even infinite. It’s called “light field” photography, and, besides eliminating failure-prone moving parts from cameras, it will let you refocus your pictures however you want, as many times as you want, any time after they’ve been captured. I don’t claim to completely understand the process, but here are what seem like the important facts:
- Instead of one big lens, “light field” cameras have lots of little “micro-lenses”—one for each pixel.
- Besides pixel color and intensity (which are recorded by conventional cameras) “light field” cameras also capture the direction of the incoming light for each pixel.
- Knowing the vector path of the light that makes each pixel allows for a sort of “reverse ray-tracing” routine that can virtually refocus the image at the software level.