135
The processor does not record the actual
focus locations of the light field. Instead,
the chip first converts the information
using an algorithm known as a
Fourier transform. The
operation winds up
with a collection of
values that have a
point-for-point equivalence to
each of the original point values
captured by the image sensor. But in a
Fourier transform, the chip can place some of
these values in a mathematical fourth dimension.5
After the transformed image has been saved to a com-
puter file, a reverse Fourier transform is applied to any collection
of values that represent any chosen plane that lies within the
light field camera’s expanded focal range. The computer can
generate several photographs from the same image values, each
bringing into focus different objects in the photo.6
The microprocessor inside the camera
receives the multiple landing locations of the rays
of light that make up thelight field, which is the
information carried by all the light entering the
camera as opposed to the thin slice of focused light in
conventional equipment. These locations represent the
patterns of focused light rays that would have converged on
the image sensor had the lens been focused nearer or farther
from the camera. (There remains a limit to how near or far the
camera can go to record extra focus information, but in even the first
version of the camera, the depth of field is expanded six times.)4
CHAPTER 8 HOW NEW TECH CHANGES PHOTOGRAPHY
Light passing through the microlenses is
not focused on one point. Instead, each
of the lenses divert the rays to a patch of
phototransistors that collect light for the
image sensor. Like the compound eyes of
an insect, the microlenses produce multi-
ple images of the light coming through
the aperture.3