ADS Technologies USBAV-708 INSTANT VIDEOMPX, Руководство пользователя

Страница: 56 / 116

The
Guide
56
Ps
yc
ho
vis
ua
l E
nha
nc
em
en
t
Wh
a
t is
p
sy
ch
ov
isu
a
l e
nh
a
nc
em
en
t?
Th
e D
C
T,
iD
C
T
a
nd
H
um
a
n
Vis
ua
l S
ys
te
m
Psychovisual Enhancement
DivX is what is known as a lossy codec—that is after encoding and decoding a
video the result will not be identical to the source, the video will have been
degraded to some extent. In other words it will have lost detail.
DivX, like other lossy codecs you may be familiar with such as MP3 or Vorbis,
uses encoding techniques based upon human perception to intelligently lose
detail where it is least likely to be noticed before degrading the remainder of the
picture. This perceptual technique forms the basis for
psychovisual
enhancement , allowing DivX to mask the artifacts inherent in MPEG4 video in
areas of the picture where error is least perceivable.
To understand how this works an explanation of the method DivX uses to
encoded the image is necessary. This is a complex subject that we will only
touch on it briefly—do not be alarmed if you do not fully comprehend it.
Inside a computers memory, images are normally represented as a two-
dimensional array of
pixels - the little colored squares that make up a complete
picture.
Storing the color value of each and every pixel is very expensive in terms of bits
and thus DivX uses a technique known as the Discrete Cosine Transformation to
convert this series of values into frequency
information. The actual process of the DCT is
an advanced mathematical topic that will not
be covered here but the main feature of the
DCT result is a set of co-efficients
representing the magnitude of the frequencies
composing the input series in order of
ascending frequency.
What is psychovisual enhancement?
The DCT, iDCT and Human Visual System