Proceed PVP - Manual (Page 8)

8
Special Design Features
Digital television (DTV) has given us greatly improved pictures and sound with
which we may entertain ourselves. However, most of us still have significant lega-
cies of existing videos and programs we like to watch that have not yet benefitted
from the performance advantages DTV offers. The PVP fills that gap by convert-
ing normal, “interlaced” video to 480p “progressive” video, a DTV standard.
But with new technology inevitably comes new terminology. A brief introduction
to the new jargon is probably in order.
Deinterlacing (“Line Doubling”) Many people mistakenly assume that a line doubler “doubles” the number of
lines in the video signal, perhaps through some sort of interpolation. Given the
name, this is a reasonable assumption. Unfortunately, it happens to be wrong.
A better name for a line doubler would be a “deinterlacer” or perhaps a “line ac-
cumulator.” The main task of a line doubler is to buffer up the odd and even
lines, reassembling them in the correct order.
Simple (read “cheap”) line doublers do just that. They put all the first field (the
odd numbered lines of information) into memory, then the second field (all the
even numbered lines), and then “reshuffle the deck” to put them back into se-
quential, progressive order. Having done so, they pass along the entire frame to
the video display. In fact, they do so twice: once for the time that each field
would otherwise have been displayed, so as to avoid flicker. This is precisely what
the unsophisticated progressive outputs built into MPEG decoder chips do.
Thus a line doubler reads out entire frames of information as often as individual
fields are normally displayed. (This is where the misleading name comes from.)
Since you are now displaying twice as much information in each unit of time, the
TV has to work faster to keep up. Specifically, normal NTSC uses a “horizontal
rate” of about 15.75 kHz. This means that the electron beam in a CRT is turning
on and off 15,750 times per second (roughly).
To display a line doubled signal, it must turn on and off twice as fast to display
all those extra pixels: about 31.5 kHz. “Regular” televisions cannot do this—this is
one of the things you pay extra for with DTV designs.
Problem: what if the interlaced camera used to shoot the video in the first place
was pointed at something that was moving? During fast action, the subject would
have moved between one field and the next. If you simply reassemble field into
frame in a simplistic way, you get jagged edges of moving objects because the
fields don’t line up. This is where motion compensation comes in. It uses sophis-
ticated analysis to compensate for such problems, ensuring that objects stay to-
gether despite the motion between one field and the next. Good line doublers
handle this sort of thing much better than simple ones do.
Another problem: video today comes from a variety of sources, including:
• film at 24 frames per second, progressively scanned during conver-
sion to video ;
• video cameras, running at an interlaced 30 frames per second;