MPEG-2 Video Coding - Digital Video and Audio Broadcasting Technology - page 111

in a 4:2:2 signal; there is thus already an irrelevance reduction in ITU-
BT.R601. It is this 270 Mbit/s signal which must be compressed to about
2...7 (15) Mbit/s in the MPEG video coding process.

7.1.1 Reducing the Quantization from 10 Bits to 8

In analog television, the rule of thumb was that when a video signal has a
signal/noise ratio, referred to white level and weighted, of more than 48
dB, the noise component is just below the threshold of perception of the
human eye. Given the appropriate drive to the A/D converter, the quantiza-
tion noise from the 8 bit resolution is already well below this threshold so
that a 10 bit resolution in Y, C B and C R is unnecessary outside the studio.
In the studio, 10 bit resolution is better because post-processing is easier
and gives better results. Reducing the data rate from 10 bits to 8 bits com-
pared with ITU BT.R601 means a reduction in the data rate of 20 % ((10-
8)/10 = 2/10 = 20 %), but this is an irrelevance reduction and the original
signal cannot be recovered in the decoding at the receiving end. According
to the rule of thumb that S/N [dB] = 6•N, the quantization noise level has
now risen by 12 dB.

Vertical blanking

625
lines

575
visible lines

Visible, active picture

Active line

Fig. 7.2. Horizontal and vertical blanking

7.1.2 Omitting the Horizontal and Vertical Blanking Intervals

The horizontal and vertical blanking intervals of a digital video signal ac-
cording to ITU BT.R601 (Fig. 7.2.) do not contain any relevant informa-
tion, not even teletext. These areas can contain supplementary data such as