The Internet Encyclopedia (Volume 3)

P1: IML/FFX P2: IML/FFX QC: IML/FFX T1: IML

Video ̇Compression ̇OLE WL040/Bidgolio-Vol I WL040-Sample.cls September 14, 2003 18:10 Char Count= 0

DIGITALVIDEOSIGNALREPRESENTATION 541

Figure 2: (a) Digital sampling of luminance and chrominance. Each cell indicates a luminance sample while chrominance

samples are indicated by cells designated by “C” and alpha channels by cells designated by “A.” The 4:4:4:4 sampling scheme

samples chrominance and alpha channels with every luminance sample. The 4:2:2, 4:1:1, and 4:4:4:4 schemes sample chromi-

nance with luminance samples while the 4:2:0 scheme samples chrominance between luminance samples. (b) Generalized

Balanced Ternary Sampling and Addressing Scheme. Seven hexagons, labeled 0 to 6 form an aggregate, which can be used to

tile the plane. The address “23” refers to the hexagon in position 3 of the aggregate in position 2. If we define the address 7 to

refer to an aggregate of hexagons, the address “077” indicates a pattern formed from the aggregation of aggregates of hexagons.

exactly 14.75 million pixels per second. Although the ITU-

R BT.601–5 Recommendation states that the duration of

the digital active line for either 525-line or 625-line video

signals is 720 samples, the active line is not clearly de-

fined. Tables 1 to 6 of SMPTE RP187–1995 recommend

values to quantify the center, aspect ratio and blanking

of video images relative to a reference image lattice of

width 1920 and height 1080 pixels, with derived pixel as-

pect ratios that Chris Pirazzi points out in his “Lurker’s

Guide to Video” to be somewhat impractical and in-

consistent with current industry practice. A “full raster”

image, coded at 10 bits per sample, contains timing,

synchronization, closed-captioning, and other ancillary

information.

Time Code

SMPTE 12M-1999 defines a time and control code for tele-

vision, audio, and film systems operating at 29.97, 25, and

24 frames per second. A video frame often contains a time

code similar to hours, minutes, and seconds as an index

for editing purposes. According to section 4 of SMPTE

12M-1999, “the exact field rate for NTSC television sys-

tems is 60/1.001 fields per second (59.94 Hz) with color

frame rate 29.97 Hz so that straightforward counting at

30 fps will yield an error of approximately+108 frames

in an hour.” “Drop frame time code” is an SMPTE time

code option that allows the time code to agree with clock

time. In drop-frame time code mode, frames 0 and 1 are

dropped on every minute with this correction negated at

the beginning of every tenth minute so that the time code

agrees with clock time as if the frame rate were truly

30 fps. Note that 25-frame systems (PAL) for which the

frame rate is exactly 50 Hz and 24-frame systems (film

media) do not require a drop frame option.

Telecine, Pull-Down, and Inverse Telecine

Teranex (Teranex, 2002) pointed out that the conversion

of 24 fps film material to 29.97 fps digital video (telecine)

takes place through a process commonly known as 2:3 (or

3:2) pull-down in which four consecutive film frames are

padded out to five digital frames of 10 digital fields. An

encoder will often try to detect that it is receiving film-

originated material and remove the redundant fields in

an inverse telecine process to provide additional compres-

sion. However, the motion and scene changes in the orig-

inal material together with the editing post-processing of

digital video can often lead to corrupt 2:3 sequences caus-

ing the encoder to spend additional resources on inverse

telecine. This can be corrected by pre-processing the ma-

terial to create a correct 2:3 cadence.

Tiling Techniques

In principle, a pixel can be any suitable shape, not just

square or rectangular. Grunbaum and Sheperd (1980)

classified the 81 isohedral tilings of the plane. Herring

(1994) of the U.S. Army Construction engineering Re-

search Laboratory pointed out that hexagonal tiles have

uniform adjacency with six nearest neighbors and may be

arranged hierarchically to establish a scalable domain-

specific software architecture model for mapping vir-

tual environments onto network computing resources.

Herring (1994) showed how a hexagonal tessellation

organizes space into a hierarchy.

Addressing Techniques

Gibson and Lucas (1982) pointed out that the tiles of the

aforementioned hexagonal scheme may be organized into

rotational hierarchies of seven hexagons and addressed

via a hierarchical place value system known as General-

ized Balanced Ternary, occasionally used in image and