🚀 go-pugleaf

We normally digitize video directly from a SIT camera, but we have a
super-VHS video cassette recorder and would like to digitize from that
sometimes.  We use a Perceptics Pixel Pipeline (Mac nuBus board) to
digitize.  I have not actually connected up the equipment and tried to
digitize from the VCR yet, but first want to understand what I am doing.

Inoue, in _Video Microscopy_ (1986) discusses using a time base corrector
to adjust the synchronization.  This is most often necessary if time lapse
recording was done (which does not apply to us).

Will someone who has digitized from a VCR please tell me if my
understanding is correct?  In particular, what controls are available on a
typical TBC?  How does a VCR generate the horizontal sync signal?  Do I
need to use a TBC?  Is there anything different about super-VHS from
regular VHS in this regard?

(1)  The VCR records one helical track for each field (about 240 visible
scan lines) of RS-170 video.  Two tracks comprise one full frame, and a
separate pulse is written in the control track for each frame.  

(2)  On playback, the VCR synthesizes vertical sync signals from the pulses
on the control track.  The control track pulses are also used to adjust the
head rotation so that the helical path which the read head follows matches
the path of the recorded track.  The tracking control adjusts this.

Apparently the VCR does not synthesize horzontal sync pulses, but simply
plays out the pulses recorded on the track.  

The head that reads the control track does not rotate, and is some distance
from the rotating head.  Because of tape dimension changes or the wrong
tape tension, the synthesized vertical sync pulses may be off and may
appear "in the middle" of a horzontal line. 

(3)  These timing inaccuracies in sync reconstruction may cause "flagging,"
which is apparent because vertical lines which should be straight appear
curved on the image, especially at the top.

(4)  Monitors have a circuit called "automatic horizontal frequency
control" which is analogous to a mechanical flywheel.  The description in
Inuoe is incomplete.  Does anyone know precisely how these work?  They have
a time constant.  Using a slow time constant works to eliminate electrical
noise.  Using a faster time constant tends to eliminate flagging.

If the VCR were synthesizing the horizontal sync, it would put it out at
precisely 15750 Hz and the AHFC time constant would have no effect on
flagging.  Thus, I conclude that the VCR does not synthesize the horzontal
sync, but must just output the sync that comes off the recorded helical
track.  The problem is that this sync may not come at the right time
relative to the vertical sync.  The monitor with its AHFC refuses to honor
the precise horizontal sync, "knowing" that it is at the "wrong" time and
rejecting it as noise.  A fast time constant AHFC would be faster in coming
to accept the "wrong" horizontal time than a slow time constant and hence
would give less flagging.

(5) The TBC digitizes several lines (or even a whole frame) of video and
then reconverts it to analog and adds sync pulses.  Presumably, it must
honor the horizontal sync pulses precisely regardless of whether they come
at the "right" time relative to the vertical pulse.  Another purpose of the
TBC is to permit synchronization of more than one VCR to permit editing. 
This requires a full frame of memory.

If you want accurate digitization, this does not sound like the way to go. 
Rather, whatever the TBC does prior to digitization should be done by the
digitizer itself.
Apparently, all the digitizer really needs to do is to start digitizing a
line at
the horizontal sync pulse rather than rejecting a horizontal sync because
it comes
at the "wrong" time relative to the vertical sync.

How to existing digitizers handle this?

--
Edward J. Huff   huff@mcclb0.med.nyu.edu   (212)998-8465
Keck Laboratory for Biomolecular Imaging
NYU Chemistry Deptartment, 31 Washington Place, New York NY 10003
Posted from the new Mac NewsWatcher 1.3d5