JVC GR-HD1 Brochure & Specs page 7

H D C y b e r C a m
performance renders 3-CCD unnecessary
Y (Luminance) information is obtained from
pairs of picture elements:
Y = (W+G) = (R+2G+B)
odd
Y = (Cy+Ye) = ((G+B)+(R+G)) =
even
(R+2G+B)
Sample Count Comparison between Single and Three CCD
1280x659 CCD(s) CCD Elements
JVC Single CCD 843,520
Similar Three CCD 843,520
The Hybrid Complementary-Primary
progressive color filter system developed by
JVC employs a filter matrix overlaying the
CCD's picture elements including two
complementary colors Yellow and Cyan (Ye
and Cy), one primary color Green (Gr), and
clear. Shifting one pixel at a time across the
CCD and combining these together in pairs
for brightness and blocks of four for color
delivers maximum resolution. This system's
superior ability to sample both Y Luminance
and RGB Chroma information results in color
performance nearly identical to using three
CCDs, proving that it is possible to deliver
Hi-Def performance with a single CCD.
C (Chroma) information is generated by
combining blocks of four picture elements:
R = (W+Ye) – (G+Cy)
G = (G+Cy+Ye) – W
B = (W+Cy) – (G+Ye)
Luma Samples RGB Chroma Samples
843,520 843,520
842,861 841,582
Aside from the number of Luminance and
RGB Chroma samples, the die-hard 3-CCD
enthusiast may wish to point out a higher
ratio of Chroma pixels to Luma pixels offered
by 3-CCD (1:1 vs 0.75:1). However, this
actually turns out to be a moot point since
4:2:0 DV or MPEG-2 encoding inevitably
reduces Chroma detail in the R–Y and B–Y
color difference components. Both record
just one sample each of color for every block
of 4 Luma samples. So instead of expending
resources on Chroma detail never used in
the recording process, JVC concentrated on
providing 3-CCD level performance with a
single CCD system tailor-made to provide all
the color sample and resolution requirements
of Hi-Def recording in MPEG-2 format.
Hi-Def MPEG-2 on MiniDV Cassette
Another remarkable aspect of this video
camera is that it records 60 minutes of HD
signals on conventional MiniDV cassette.
The benefits are obvious: MiniDV
cassettes, "the defacto standard",
are widely available and affordable
compared to other formats used in
professional systems, and by using
a common cassette mechanism with
DV, it was possible for the GR-HD1 to
offer regular DV recording as well.
Digital MPEG-2 Hi-Def and SD Recording
While the picture is encoded one frame at a
time for DV, the time axis is encoded as well
for MPEG-2, conforming to the HD Digital
VCR Conference's Part 7 (DVB) and Part 8
(ATV) standards. The GR-HD1 uses a GOP
(Group Of Pictures) of 6 for HD and 12 for SD,
with a frame order of IBBPBB. The "I frame" is
MPEG-2
In HDTV, high power MPEG
encoding eliminates redundant
information over time.
I
time
I : I Frame (Intra Coded Frame)
B : B Frame (Bi-directionally
Predictive Coded Frame)
P : P Frame (Predictive Frame)
High-Speed Search Playback Possible
The "I frame" can be played back without
using any other frame and is recorded in
the appropriate variable-speed playback
*The difference between HDTV TV lines and regular TV lines.
All "TV line" standard measurements are relative to the diameter of a circle the same size as the
picture height. HDTV 16:9 Horizontal TV lines figures appear to be 75% of what the equivalent
4:3 Horizontal TV Lines would be if the measurement was relative to the whole picture width,
since the image is wider. (For example, for the GR-HD1 HD mode would be around 1244 TV
lines per picture width, while the count would be about 933 TV lines per picture width for a 4:3
camera if both had 700 TV lines counting by standard horizontal resolution). DV has a practical
recorded horizontal limit of 540 4:3 TV lines, 1280x720p HD has a practical recorded limit of 720
16:9 TV lines using the standard measurement method.
This was made possible since the "Hi-Def
MPEG-2 on MiniDV" encoder and
decoder developed by JVC together with
NTT records Hi-Def Component MPEG2-
TS signals in the same track format as DV.
Tape low edge (Reference edge)
a complete frame without reference to any
past or future frames; the "P frame" is
encoded relative to the past reference frame;
and the "B frame" is encoded relative to the
past reference frame, future reference frame,
or both.
B B P B B I
GOP = 6
area on the tape, allowing high-speed
search during playback.