Shaped video – Grass Valley Switcher Products User Manual
Page 134

134
Switcher Products — Protocols Manual
Section 6 — Still Store Image File Format
For SD formats CCIR 601 defines the basic video signal levels used by the
still store and the conversion between RGB and Y, Cr, Cb as:
•
Y = 0.299R + 0.587G + 0.114B
•
Cr = 0.713 (R - Y) = 0.500R - 0.419G - 0.081B
•
Cb = 0.564 (B - Y) = 0.500B - 0.169R - 0.331G
For HD formats SMPTE 274M (1920 x 1080) and 296M (1280 x 720) specify:
•
Y = 0.2126R + 0.7152G + 0.0722B
•
Cr = (0.5/(1 - 0.2126)) * (R - Y)
•
Cb = (0.5/(1 - 0.0722) * (B - Y)
In addition, headroom is specified for overshoot and undershoot.
•
Black is a Y code word of 64
•
White is a Y code word of 940
•
Cr and Cb sit on a pedestal of 512 to avoid negative numbers and has a
range of ± 448 about this central value.
These factors need to be taken into account when scaling values.
Cr and Cb are half the bandwidth of Y. Proper filtering should be applied
during RGB to Y, Cr, Cb conversion to avoid aliasing. Half the Cr and Cb
samples are then discarded. Cr and Cb are both derived from the same
sample (first, third, fifth, etc.).
In addition, code words below 4 and above 1019 are reserved for sync and
control words. The Y, Cr, and Cb values must be limited accordingly.
While CCIR 601, SMPTE 274M and SMPTE 296M specifies pixels per line
and the number of lines per field, SMPTE RP-187 does a better job of
defining things. RP-187 specifies images in terms of a “clean aperture”
(what the end image will be cropped to) and a “production aperture.” The
still store stores the entire production aperture. However, cropping and
positioning is done in “screen units” (also defined in RP-187) that are based
on clean aperture. Note that screen units are different for 4:3 and 16:9 aspect
ratios.
Shaped Video
Modern video production equipment usually supports two mechanisms of
transferring video between equipment.
The most common approach used in all digital environments is known as
“shaped video” because the video produced by one device has already
been modified by its accompanying key. The receiving device “cuts a hole”
in the background image and adds the incoming video to it without further
processing.