(^116) PART 2 HOW DIGITAL CAMERAS CAPTURE IMAGES
How a Digital Camera Squeezes
Video Down to Size
Compressing one picture on-the-fly is daunting enough for a digital still camera. But consider the task that
confronts a video camera: compressing 15 to 30 images each second—and putting the squeeze on sound
at the same time. There is no time for lossless compression, even if it could do a good enough job to store
even a short feature on a DVD. The technology universally used for pro-
fessional and amateur video is a lossy compression called MPEG,
named after the organization that created it: the Motion Picture Experts
Group. MPEG compression is likely to be used on anything you watch
on TV, your computer, or iPod.
An essential part of video compression strategy is to limit the
amount of work the video camera has to do to record so many
frames each second the camera’s rolling. Standard TV
(non–high definition with a 4:3 aspect ratio) is made up of
720 pixels across and 576 pixels ver-
tically. Depending on the sophistica-
tion of a digital video camera, frames
may be limited to 640 pixels wide
and 480 pixels high, or even as few
as 352 ×288. At 640 ×480, the
camera is relieved of the task of
recording information for a hundred
thousand pixels for every frame it
processes.
Inside a video camera, you’ll find a chip
that specializes in MPEG compression
(or circuitry that does the same job built
into the master microproces-
sor). Part of the job MPEG
performs is similar to the lossy
JPEG compression used to
shrink still photos. MPEG looks
for redundant data. The wall in
the background of a video may
be painted a solid red, making
it unnecessary for a video
camera to record the same
information about each of the
red pixels that make up the
wall. As JPEG compression did
with a blue sky in the example
a few pages previous, MPEG
establishes reference values for
the wall’s shade of red and
refers all the pixels that make
up the wall to that color.
2
1
singke
(singke)
#1