virtual car. Gen locking [OT][EE]
Robert Rolf email (remove spam text)
Richard Ottosen wrote:
> >From what I understand, the color locking is the hard part. I was once
If you want to dissolve between two camera signals, yes, it is difficult
to get the phase lock down to the sub degree range required.
> told that half of the circuitry in the color cameras I borrowed was the
> video locking function. This may no longer be true with the changes in
An exaggeration. I've been in video for 20 years. While substantial
is needed for broadcast grade locking (10% at most) your alternating
field/frame approach does not require it. Gen locking adds about $50 to
cost of a $300 CCD color camera. A lot of the chip sets support it
directly, just add the jack and coupling cap/resistors.
The funny thing is, the new broadcast switchers don't need 'timed'
camera sources anymore. It's all done digitally now, so ANY source
(including sloppy VCR's) will do.
> One change in technology I need to explore is inexpensive time base
> correctors. With picture-in-picture built into TV's and frame capture
> cards for PC's I expect it would not be that hard to make a simple TBC.
It isn't. $150 per channel for a dual channel TBC for a PC (DPC systems
in Montreal?). Just look in any of the consumer 'Desk Top Video'
for adverts. Signetics/Philips makes several chips that will do the job
Ironically a PIP video card is already doing 'frame synching' (which is
actually a lot harder than simple Time Base Correction). The kick is
that the VGA card is running 60-80FPS non interlaced vs NTSC's 30 with
Also realize that a cheap TBC many only provide a few (5-10) lines of
'correction' and will not have a enough range to 'frame lock' unless
it explicitly says that it is a 'full frame' TBC.
You realize that you only need ONE frame synchronizer to do what you
want? You use one camera as the reference that the other is locked to.
HOWEVER, having done this, I found that one could perceive the slight
difference in timing between the two eyes when the cameras were quite
out of sync. True NTSC is a 4 field sequence so you could see up to 67mS
of delay between the two eye views.
You really DO want to lock the _camera_ scans to remove this 'headache'
[literally]. If you don't, fast pans will make the image jump in and
out in depth as one eye sees the same image 67 mS before the other.
In fact, this is exactly the technique used by the infamous super bowl
3D of several years ago. The darkened eye has a slower visual response
and so fast left right object motion gave the illusion of depth.
And for this exploit the perpetrators got a million bucks?
> Obviously some pretty fancy chips exist that could make it possible. Oh,
> so many projects and so little time ;-)
I know EXACTLY what you mean.
BTW the NASA technical channel (Tr5, G2 85dW) does time division
multiplexing of the various cameras on the cape in the hours prior to
A digital VCR with 'strobe' function can usually separate out the
two streams into something more viewable. It's awesome being inches away
from the external fuel tank as they do a systematic inspection of it's
surface. It's also pretty cool to be looking over the astronauts
as they work.
NASA tv is available on the web as Realvideo or Mediaplayer formats by
drilling down from:
See also: www.piclist.com/techref/index.htm?key=
You must be a member of the
piclist mailing list
(not only a www.piclist.com member) to post to the