[sci.electronics] "EIA Industrial Electronics Tentative Standard No. 1"

poynton@vector.Sun.COM (Charles A. Poynton) (08/11/89)

Isaac S Wingfield <ISW@cup.portal.com> asks:

> *Exactly* what are the "equalizing pulses" for?
>
> No guesses, please; If you *know*, please enlighten me.

Indeed as Bob Myers suggests [no References: line, otherwise I'd quote],
the equalizing pulses have to do with interlace.  Since you ask for an
*exact* explanation, I'll explain in detail.

The first thing you need to know is that if you attempt to just jam
composite video down to -40 IEEE solidly for three lines of vertical sync,
the horizontal oscillator will freerun for those lines, and it will take a
few (or many) lines to get synchronized back to the incoming video when H
syncs resume.  This used to be known as "EIAJ" or "industrial" sync, and
can lead to 'flag waving' at the top of the picture, in the worst case.
So it is necessary to 'serrate' the vertical pulse, in order to provide H
sync information during vertical.

The next thing you need to know is that in an interlaced system, the first
vertical (broad) sync pulse in one field needs to be coincident with an H
sync, and the first vertical pulse in the other field needs to occur half
way between two H syncs.  If you don't kick V very very nearly half way
through a line, at the top of the even field, then the lines 'pair', that
is bunch up together leaving black bar in between, and this causes an
objectionable picture im-pair-ment (pardon me, it's getting late).

Now, take yourself back to 1941.  You've got vacuum tubes, but not very
many of them, and you've got Rs and Cs.  To separate the vertical and
horizontal components of sync, you use an R and a C in each leg, as cheapo
HP and LP filters.  Basically the V filter is an integrator that ramps up
while composite video is below -20 IEEE (whoops, IRE, this is 1941), and
leaks back down slowly when it's above.  During the picture lines, the
leak rate is faster than the little pumps it gets from the little 4.7 us H
syncs, and the integrator output sits down at some saturation level, but
when broad pulses start arriving, with about 54 us of time below -20 IRE
per line, it ramps up considerably.  There's a slicer that kicks the V
flyback when the integrator passes a threshold.  The trouble is, because
the filter pair is so cheap, a little H contribution gets into the
separated V.  The amount and timing of that extra energy depends on which
field you're in, and if there were just full-width H syncs and a line-wide
V sync, that would induce line pairing.

So equalization pulses were introduced, so that if you look at any V sync
event, a string of six broad pulses, the immediate neighborhood before and
after comprises exactly the same pattern:  six 2.3 us "equalizing" pulses
at twice H rate, and six 27 us broad pulses at twice H rate.  That makes
one field look "equal" to the other in terms of H sync contribution to the
V integrator; in other words, a thresholded V sync event can't tell which
field it came from.  This arrangement prevents line pairing.

The resulting equalization/broad pulse pattern is what produces the
sideways-"T" pattern that you see if you turn up the brightness and roll
down vertical hold (if anybody's still GOT a vertical hold control).

Now of course, with decent analog and digital ICs, all of this is quite
redundant.  Work done in the Soviet Union a few years ago concluded that
we could now safely do away with all but a single pre-equalizer even for
the current aging population of 525/59.94 and 625/50 receivers.  HDTV will
retain twice-line-rate broad pulses, but have a single pre-equalizing
pulse in one field and a single post-equalizing pulse in the other.

By the way, Bob, I insist on "EIA No. 1" as part of my campaign to
embarrass the EIA and the SMPTE into adopting a standard.

C.

-----
Charles A. Poynton			Sun Microsystems Inc.
<poynton@sun.com>			2550 Garcia Avenue, MS 8-04
415-336-7846				Mountain View, CA 94043

"Japan has no laws against damage to its flag, but it has strict laws
forbidding the burning of foreign flags lest this give offense to the
country in question." -- The Economist, July 1, 1989, p. 19.
-----