Apple Final Cut Express HD - Manual (Page 1040)

1040 Part XIII Appendixes
Video Compression
Once a video signal is digital, it requires a large amount of storage space and
transmission bandwidth. To reduce the amount of data, several strategies are employed
to compress the information without negatively affecting the quality of the image.
Some methods are lossless, meaning that no data is lost, but most are lossy, meaning
that information is thrown away that can’t be retrieved.
Some simple methods of data compression are:
 Throw away pixels at regular intervals: This essentially scales the image, or makes it
more blocky.
 Average several pixel values together: This involves taking several adjacent pixel values
and averaging them together, resulting in a single rectangular pixel that approximates
the value of several. For more information, see “
Pixel Aspect Ratio ” on page 1033.
 Throw away color channel information at regular intervals: This results in color sampling
ratios like 4:2:2 and 4:1:1. Ideally, throwing away this color information is not noticeable
to the viewer, but may be a problem if you are trying to do detailed color correction or
chroma-keying that requires a lot of color information to start with.
Lossless Codecs
Once these basic methods have been employed, much more intensive algorithms can
be employed to reduce the amount of transmitted and stored image data.
Mathematical algorithms can be used to encode and decode each video frame. These
codecs (enCode, Decode) must be installed in the VTR or software you are using to play
back your video. For example, QuickTime supports many different video codecs for
video export and playback.
The simplest encoding algorithm, called run-length encoding, represents strings of
redundant values as a single value and a multiplier. For example, consider the following
bit values:
0000000000000000000000001111111111111111000000000000000000000000
Using run-length encoding on the bit values above can reduce the amount of
information to:
0 x 24, 1 x 16, 0 x 24
Or in binary:
0 [11000], 1 [10000], 0 [11000]
In the example above, the original 64 bits can be transmitted using only 18 bits.